Cost-Effectiveness Threshold for Healthcare: Justification and Quantification

Abstract

Every public health expenditure, including the one that saves lives or extends life expectancy of particular persons (target population), bears a cost. Although cost-effectiveness analysis (CEA) is routinely performed in health policy, ethical justification of CEA is rarely discussed. Also, there is neither consensus value nor even consensus method for determining cost-effectiveness threshold (CET) for life-extending measures. In this study, we performed ethical analysis of CEA by policy impact assessment based on connection of health and wealth (poorer people have statistically shorter life expectancies) and concluded that CEA is not only a practical but also an ethical necessity. To quantify CET, we used three independent methods: (1) literature survey of analyzing salaries in risky occupations, (2) utilizing Prospect Theory suggesting that people value their lives in monetary terms twice more than their lifetime earnings, and (3) literature survey of the U.S. current legal practice. To the best of our knowledge, nobody applied method (2) to determine CET. The three methods yielded rather similar results with CET about 1.0 ± 0.4 gross domestic product per capita (GDPpc) per quality-adjusted life-year. Therefore, a sum of not higher than 140% GDPpc is statistically sufficient to "purchase" an additional year of life-or, alternatively, to "rob" one year of life if taken away. Therefore, 140% GDP per capita per quality-adjusted life-year should be considered as the upper limit of prudent and ethically justified expenditure on life extension programs.

Full text

Original Research Article
INQUIRY: The Journal of Health Care
Organization, Provision, and Financing
Volume 59: 1–8
© The Author(s) 2022
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/00469580221081438
journals.sagepub.com/home/inq
Cost-Effectiveness Threshold for
Healthcare: Justification and Quantification
Moshe Yanovskiy, PhD
1
, Ori N. Levy, PhD
1,2
, Yair Y. Shaki, PhD
1
,
Avi Zigdon, MHA, PhD
2,3
, and Yehoshua Socol, PhD
4

Abstract
Every public health expenditure, including the one that saves lives or extends life expectancy of particular persons (target
population), bears a cost. Although cost-effectiveness analysis (CEA) is routinely performed in health policy, ethical justification of
CEA is rarely discussed. Also, there is neither consensus value nor even consensus method for determining cost-effectiveness
threshold (CET) for life-extending measures. In this study, we performed ethical analysis of CEA by policy impact assessment based on
connection of health and wealth (poorer people have statistically shorter life expectancies) and concluded that CEA is not only a
practical but also an ethical necessity. To quantify CET, we used three independent methods: (1) literature survey of analyzing salaries
in risky occupations, (2) utilizing Prospect Theory suggesting that people value their lives in monetary terms twice more than their
lifetime earnings, and (3) literature survey of the U.S. current legal practice. To the best of our knowledge, nobody applied method (2)
to determine CET. The three methods yielded rather similar results with CET about 1.0 ± 0.4 gross domestic product per capita
(GDPpc) per quality-adjusted life-year. Therefore, a sum of not higher than 140% GDPpc is statistically sufficient to “purchase”an
additional year of life—or, alternatively, to “rob”one year of life if taken away. Therefore, 140% GDP per capita per quality-adjusted
life-year should be considered as the upper limit of prudent and ethically justified expenditure on life extension programs.
Highlights
What do We Already Know About This Topic?
Cost-effectiveness analysis (CEA) is routinely performed in health policy. However, very different methods for determining
cost-effectiveness threshold (CET) and very different values for CET can be found in the literature. Also, ethical justification
of CEA is rarely discussed.
How does your research contribute to the field?
Excessive healthcare spending claims more lives than it saves. The value of 140% of gross-domestic-product per capita per quality-
adjusted life-year is the upper limit for prudent expenditure on healthcare and safety. This important side effect was mainly ignored in
decisions on lockdowns targeted to save lives.
1
Department of Industrial Engineering, Jerusalem College of Technology, Jerusalem, Israel
2
Disaster Research Center, IL, Ariel University, Ariel, Israel
3
Department of Health Systems Management, School of Health and Medical Sciences, Ariel University, Ariel, Israel
4
Department of Electrical and Electronics Engineering, Jerusalem College of Technology, Jerusalem, Israel
Corresponding Authors:
Yehoshua Socol, Jerusalem College of Technology, Havaad Haleumi 21, Jerusalem 91160, Israel.
Email: socol@jct.ac.il
Avi Zigdon, Department of Health Systems Management, School of Health and Medical Sciences, Ariel University, Ariel 40700, Israel and Disaster Research
Center IL, Ariel University, Ramat HaGolan 65, Ariel 40700, Israel.
Email: aviz@ariel.ac.il
Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons
Attribution-NonCommercial 4.0 License (https://creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use,
reproduction and distribution of the work without further permission provided the originalwork is attributed as specified on the SAGE and
Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).

Keywords
health policy, ethics, risk management, cost-benefit analysis, willingness to pay
Introduction
Investment in healthcare, technological safety measures,
safety and environment-protection regulation decrease risks
and extend lives, yet increase the economic burden.
1
Thus,
resource restrictions are expected, and countries will be
forced to prioritize medical investments.
2
Associating human life with monetary value is psycho-
logically difficult. The opposition to a cost-effectiveness
analysis (CEA) in health and safety decision-making is not
unfounded,
3
but CEA still unavoidable.
4-6
CEA is routinely
performed in health policy, though decisions are rarely, if
ever, made based on cost-effectiveness only.
6
Seventeen
countries at least have formal cost-effectiveness thresholds
(CETs).
7
Not only health policymakers but also medical
practitioners can no longer consider only benefits and side
effects but should also be aware of the associated cost.
8
The terms “cost-effectiveness analysis”and “cost-benefit
analysis”are close. However, cost-benefit analysis places
monetary values on health outcomes or life, and therefore
raises many ethical objections.
8
Ethical justification of CEA for life-saving and life-
extending measures is rarely discussed—it is usually stated
that each dollar invested in a potentially life-saving program
(e.g., cancer prevention) is not invested in another potentially
life-saving program (e.g., cancer treatment). However, it does
not address the frequent claim of ALARA (As Low as
Reasonably Achievable) principal proponents: money spent
on healthcare and safety measures (e.g., by private agents due
to safety regulation) would not be invested in alternative life-
extending projects.
9
So, the ethical justification of cost-
effectiveness analysis remains an open question.
Regarding value of cost-effectiveness threshold, there is
neither consensus value nor even consensus method for de-
termining the latter. Essentially, two alternative approaches—
welfarist and extra-welfarist—are commonly used.
7
According to the extra-welfarist approach, in order to
estimate CET, one would ideally consider all the reasonable
factors—taking risk, safety of products, diet etc. The term “value
of statistical life”(VSL)
5
is in official use by the U.S.
government.
10,11
(Although the term VSL probably raises neg-
ative connotation of equating human life to some monetary value,
its meaning is just that there is human life cost for any spending.)
Determining CET via VSL follows the extra-welfarist approach
when the government is supposed to quantitatively assess public
expenditure that is worth being invested in potentially life
expectancy-extending policies and actions.
12
However, this task
is formidable due to multiple uncertainties.
Alternatively, the welfarist approach is probably the most
straightforward practical approach to determine CET by esti-
mating “willingness to pay”(WTP)—how people themselves
value their life and health in monetary terms.
12
In our opinion, the
welfarist approach should be preferred, and the term WTP itself
describes most exactly the mechanism behind the assumption that
the state should not provide citizens with services (including life-
extension services) that are less cost-effective than the citizens
themselves are willing to pay for such services.
13
In this paper, we address first the question of ethical
justification of cost-effectiveness analysis. Then, we review
the existing methods of CET determination and point on their
limitations. Finally, we apply three different methods, all of
them based on WTP approach, to estimate justified cost-
effectiveness threshold for life extension spendings.
Methods
To justify ethically the applicability of cost-effectiveness
analysis to healthcare, we performed policy impact assess-
ment based on connection of health and wealth. Since poorer
people have statistically shorter life expectancies, the
ALARA approach to risk has an inherent ethical problem: by
statistically extending life expectancy of some target pop-
ulation, we statistically reduce life expectancy of the others.
For the quantification of cost-effectiveness threshold, we
used three independent methods, all of them based on WTP
approach, and performed literature survey of each of them:
(1) analyzing salaries in risky occupations, (2) using Prospect
Theory assuming that people value their lives twice more than
their lifetime earnings, and (3) comparing with the U.S.
current legal practice. To the best of our knowledge, nobody
applied method (2) to cost-effectiveness analysis.
Results
Ethical Justification of Cost-Effectiveness Analysis for
Healthcare and Safety
There is an approach that since the value of human life is the
highest possible from a societal perspective, any risk should
be kept “as low as reasonably achievable”—the ALARA
principle,
9
no matter the price, if it is bearable. However, the
ALARA approach has an inherent ethical problem; by
What Are Your Research’s Implications Towards Theory, Practice, or Policy?
Every program or policy with cost exceeding 140% of gross-domestic-product per capita per quality-adjusted life-year does not pass
the criteria of prudent and ethically justified expenditure.
2INQUIRY

statistically extending life expectancy of some target pop-
ulation, the society statistically reduces life expectancy of the
others.
The reason is that well-being and the life expectancy of the
modern industrialized society is based on its material basis.
Considering an unreasonable extreme, with all resources
invested in healthcare, people would just die of hunger.
Another extreme—nobody suggests giving up motorized
transport though it claims yearly about 35,000 lives in the
U.S. alone. In general, each dollar collected as tax (even to be
invested in a potentially life-saving program) makes the
taxpayer poorer by one dollar. Each dollar spent by a com-
pany due to safety regulation decreases by one dollar its
profit, with direct impact on the workers’salaries and
shareholders’income—also making them poorer. And sta-
tistically, poorer people have shorter life expectancies.
There are many mechanisms that contribute to the cor-
relation between socio-economic status and life expectancy.
Poorer people take higher-risk jobs, they work extra hours
stressing their health, they use cheaper, and consequently less
safe, products, their lifestyle is less healthy, their diet is less
healthy, their health insurance is less comprehensive etc. For
example, it has been shown that demand for health insurance
is highly elastic
14
—that is, people purchase health coverage
plans proportionally to their income. Personal experience of
one of the authors (YS) illustrates just one of multiple
mechanisms leading to lower life expectancy of poorer
people: Two of his neighbors lost their lives in two unrelated
traffic accidents. Both drove very cheap cars because of their
economic status, and both would most probably survive with
minor injuries—based on the post factum analysis of the
accidents—would each of their cars be just about US$8000
more expensive. The corresponding loss of life in the two
cases was 30 years (50-year-old woman) and 50 years (30-
year-old man)—about two hundred US$ (!) per life-year on
average.
Not every drop in income significantly affects health. If the
drop is slight, the expenditures on food, medicine, and health
insurance are usually not affected. The family also tries to
preserve the usual vacation patterns, cutting off other
spendings. Moreover, if a household/person reduces spend-
ings that have little bearing on health or a negative bearing
(e.g., purchase of alcohol, tobacco, or less healthy restaurant
food), no negative effect occurs. As an anecdotal example one
can remind that the food shortage during the Second World
War led to considerable improvement in health of many celiac
patients (and ultimately led to the understanding of the celiac
mechanism).
However, such a stepwise (rather than linear) health effect
of income reduction at the microlevel (households) does not
mean that a small additional burden placed on the society (e.g.,
a tax) does not pose risks to health and life. Small changes in
national income mean significant changes in the expenses of
particular households.
15
For example, even a small increase in
company tax can bankrupt a company balancing on the brink
of survival. And unemployment or a significant reduction in
business income will undoubtedly force most affected
households to reduce their costs, affecting their health.
It should be emphasized that healthcare costs in most
developed countries are already a heavy burden on national
economy
16
:8–12% of GDP in OECD countries,
17
almost
17% in the USA.
18
Not accidentally, shifting a substantial
fraction of health expenditures from the medical system to
socio-economic targets has been proposed.
19
It is very likely
that easing safety regulations and reducing healthcare pro-
grams of questionable efficiency will save more life years
than will be lost without them.
Therefore, every public expenditure, including saving
lives or extending life expectancy of particular persons, has
unwanted but unavoidable side effect of statistical shortening
of life expectancy of other people.
The above analysis shows that a policy with cost-
effectiveness below some threshold claims more life than
it saves. The question of decision-making is therefore
whether the direct effect, or policy goal (life expectancy
extension for the target population), is stronger than the side
effect (life expectancy shortening for non-target population).
If the side effect is stronger than the direct (policy goal), the
net result of such policy is the statistical shortening of life.
20
Such situation has been even characterized by the term
“statistical murder.”
20
Therefore, cost-effectiveness analysis
is not only practical but also ethical necessity.
Cost-Effectiveness Threshold (CET) Quantification
Analysis of the existing methods. The method of determination
of CET via WTP follows the welfarist approach. The wel-
farist approach is probably the most straightforward practical
approach to determine CET by estimating how people
themselves value their lives and health in monetary terms.
12
In our opinion, the welfarist approach should be preferred
over extra-welfarist, and the term WTP itself describes most
exactly the mechanism behind the assumption that the state
should not provide citizens with services (including life-
extension services) that are less cost-effective than the citi-
zens themselves are willing to pay for such services.
13
We are
going to discuss this mechanism now.
Although accepting certain (or even very probable) death
for money is unacceptable, taking small risks for money is
routine: every profession is associated with some risk and
there are professions that are riskier than the others (fire-
fighters, police, construction workers etc.). If on average
people are ready to take risk of death with, for example,
probability 1/1000 (one of a thousand) for $1500, then WTP
should be estimated as $1.5 million. Namely, to get $1.5
million of earnings, 1000 people on average will take the risk
1/1000 to die, and one on average will die. Though people
take their risks voluntary, the net effect is that public ex-
penditure of one WTP statistically claims one human life. The
above method of valuing life was proposed by Adam Smith
Yanovskiy et al. 3

more than two centuries ago and has been used since then in
economic analysis as well as in legal practice.
21
Therefore,
reasonable value of CET per statistical life saved should be
set about the value of WTP. It is not about monetary value of
life. It is about extending life population-wide.
As mentioned above, despite a growing body of empirical
studies on CET, no consensus has emerged regarding the
methodology. The World Health Organization (WHO) has
recommended CET equaling factor 1–3 times the gross do-
mestic product (GDP) per capita.
22
Marseille et al stressed,
however, that while willingness to pay for health care is related
to income, this relationship maybe nonlinear, so the CET as a
fraction of GDP may be too stringent in high-income countries
(rejecting some efficient options) and too lax in low-income
countries (accepting some inefficient options).
3
British researchers
23
estimated CETs based on recent em-
pirical estimates of opportunity costs from the UK National
Health Service. The method of opportunity costs perfectly fits
countries with purely-market economy. However, in all countries
with reliable statistics healthcare services are heavily subsidized
and strictly regulated by the state. The opportunity costs are
therefore influenced by the government involvement which
differs significantly in various countries, so even surveys like
Shiroiwa et al
24
cannot provide a reliable basis for comparison.
CET Value Estimation
In light of the complexities described in the previous section,
we have chosen to implement a less sophisticated and hopefully
more practical approach. Our analysis is based on three inde-
pendent estimations: (1) analysis of salaries in risky occupations,
(2) analysis based on Prospect Theory (Nobel Prize in Eco-
nomics for 2002 to Daniel Kahneman) assuming that people
value their lives twice more than their lifetime earnings, and (3)
comparison with the current U.S. legal practice.
1. In 1976, Thaler (later awarded the Nobel Prize in
Economics for 2017) and Rosen analyzed salaries in
different occupations and compared the salaries with
risk (mortality). They estimated VSL to be $200,000 ±
60,000 in 1967 dollars.
25
In 2019 dollars, the above
estimation corresponds to $1.53M ± 0.46M based on
Consumer Prices Index.
26
In order to perform cost-effectiveness analysis of evacu-
ation we need not only VSL, but also value of spending
equivalent to statistical loss of one life-year. The latter can be
obtained by dividing VSL by half of life expectancy at
birth—since statistical (accidental) death can occur randomly
at any time during the lifespan. For life expectancy at birth of
about 80 years, typical for the developed countries, VSL
should be divided by 40 years. The above estimation yields
therefore CET = $38,250 ± 11,500 per life-year (the accuracy
of the numbers, here and below, is certainly spurious; we keep
this spurious accuracy till the final averaging).
2. Thaler’s estimation of VSL fits the results of Prospect
Theory
27
that people quantify potential loss approx-
imately twice as much as potential earning.
28
Namely,
people quantify losing $10, for example, as much as
not gaining $20. The conclusion of 2:1 ratio in
preference between loss and gain was made based on a
vast body of psychological experiments; it was verified
by analysis of real-life decision-making, for example,
comparing purchasing new insurance with renewing
existing. According to Prospect Theory, this 2:1 rule
applies not only to money but also to other goods like
vacation days etc. In our case, we assume that people
value their lives (that they have and can lose) in monetary
terms approximately twice as much as the anticipated
earnings. To the best of our knowledge, such estimation
of CET has not been performed before.
Let us perform the corresponding calculation. Net yearly
after-tax income of an average U.S. worker was about $43,000
in 2019 (https://taxfoundation.org/us-tax-burden-on-labor-
2020/). Doubling this sum yields CET = $86,000 per life-year.
3. In the U.S. legal practice, the median settlement com-
pensation for the victims of the 9–11 attack was $1.7M (in
2017 dollars), and median death compensation awarded by
jury in 2009–2013 was $2.2M (also in 2017 dollars).
29
We
can therefore take $1.95M ± 0.25M for the legal practice
with corresponding CET estimation of $48,750 ± 6250. In
2019 US$, CET = $50,800 ± 6520.
It is worth explaining why legal practice is relevant to the
determining of CET. Judicial practices in a country with a
respected court accumulate a large volume of practical
decisions in various situations. The existing judicial
practice, which does not cause condemnation or at least
wide discussion, gives, therefore, a good assessment of the
public acceptability of the decision from the point of view
of society. Therefore, the assessment of the “value of human
life”in court to determine compensation provides important
guidance for CET/VLS/WTP values acceptable by the
society.
21
The CET values are summarized in Table 1. The difference
between the three values is about 2-fold, which seems to us
rather modest taking into account the very different esti-
mation methods.
Rigorous statistical averaging the results of the three
methods cannot be done due to many unknowns. Instead of
inventing mathematical method of questionable applicability,
we just take the CET value as the mean value of the three
estimations, and the CET range as the range of the values
from US$ 38 250 to US$ 86 000. Rounding the numbers to a
reasonable accuracy yields:
CET = 60,000 ± 25,000 US$ per quality-adjusted life-year.
Our estimation was performed on the U.S. data.Wecan
generalize by expressing the result via gross domestic product
4INQUIRY

(GDP) per capita—US$ 65,000 in 2019.
30
Rounding to
reasonable accuracy yields:
CET = 1.0 ± 0.4 GDP per capita (GDPpc) per life-year.
We find therefore that in the U.S., both people as indi-
viduals and the society in general value their lives at about 1.0
± 0.4 GDPpc per QALY. Therefore, as discussed in the
section on ethical justification, a sum of not higher than 1.4
GDPpc is statistically sufficient to “purchase”an additional
quality-adjusted year of life—or, alternatively, to “rob”one
year if taken away. So, 1.4×GDPpc/QALY should be con-
sidered as the upper limit of prudent expenditure on
healthcare and safety: higher expenditure most probably
claims more life than it saves.
The upper limit for prudent expenditure on healthcare and
safety—140% GDP per capita per quality-adjusted life-year—
is the single most important result of this study.
Discussion
Though the value of human life is the highest value for the
society, life extension by means of healthcare and safety
should not become a super-goal consuming all the rea-
sonably available resources. The reason for this is that
every public expenditure, including saving lives or ex-
tending life expectancy of particular persons (target pop-
ulation), has unwanted but unavoidable side effect of
statistical shortening of life expectancy of non-target
population.
It has been stated that the impact of socio-economic factors
on health is enormous compared to the power of healthcare to
counteract these factors. A metaphor for the connection of
socio-economic and health parameters is the “New York
subway map”
31
: From Manhattan to the South Bronx, life
expectancy declines by 10 years, half a year for every minute
on the subway. No medical intervention, either existing or
even conceivable, has the same order of magnitude of effect
on health.
Our value for CET was derived by three independent
methods considering different sides of economic equilibrium:
(1) by analyzing salaries in risky occupations,
25
(2) by as-
suming that people value their lives twice more than the
wealth they earn (Prospect Theory),
27
and (3) by comparing
with the U.S. current legal practice.
29
In microeconomics
terms, the former method (1) is based on demand-supply
equilibrium in the labor market. The Prospect Theory method
is based on supply-side considerations determining readiness
of agents (employees) to take risks. The latter method (3) is
based on numbers manifesting the readiness of the society to
reimburse risk posthumously.
The calculated value is in excellent correspondence with
the values accepted in the field of public healthcare. It has
been estimated that the U.S. national health insurance is
associated with an average cost-effectiveness of about US$
50,000 per QALY (quality-adjusted life year) gained.
13
It is
true that values used for decision-making often have an upper
limit in the $100,000 to $150,000/QALY or higher, as
mentioned by Birch and Gafni
32
or Padula et al.
33
However,
Padula et al
33
note that there is a strong association between
estimated ICER (incremental cost-effectiveness ratio) values
and chosen CET: the regression analysis indicated that CETs
have a baseline value of $52,000 and grow by $0.37 for each
dollar increase in the estimated ICER. Therefore, higher CET
values seem to be biased.
In the United Kingdom, National Institute for Health
and Clinical Excellence (NICE) adopted a cost-
effectiveness threshold range of $40,000 to $60,000 per
QALY.
34
The review of Bonis and Wong
8
estimates
healthcare CET as US$50,000 to $100,000 per QALY.
Cameron et al estimated CETs in 17 countries and found
CET/QALY values in units of GDP per capita varying
widely among different countries, from as low as 0.28
(Thailand) to as high as 4.2 (Belgium).
7
This being said, the
CET values seem to be highly correlated with GDP—see
Figure 4 of Cameron et al.
7
For countries with GDPpc below
about $35K, CET/GDP is about one. For higher-income
countries, CET/GDPpc varies from slightly above one (Swe-
den) to 4.2 (Belgium). It is important to mention that no sig-
nificant correlation between CET/GDP ratio and life expectancy
has been observed (Figure 3 of Cameron et al
7
). For example,
Sweden with the lowest CET/GDP ratio among the high-income
countries has longer health-adjusted life expectancy than Bel-
gium with the highest CET/GDP.
Numbers very different from the calculated in this work
can be found in literature; different government agencies
estimate VSL (value of statistical life) to justify their policies.
Their estimates should be viewed with extreme caution.
First, methodologically there are multiple sources of bias,
usually towards overestimating VSL.
11
Table 1. Independent estimations of cost-effectiveness threshold.
Method
Cost-Effectiveness Threshold (CET), US$ per
Quality-Adjusted Life-year
1 Analysis of salaries in risky occupations
26
38,250 ± 11,500
2 Based on the prospect theory (people value their lives twice
more than the lifetime earnings
28
)
86,000
3 Comparison with the current U.S. legal practice
30
50,800 ± 6520
TOTAL 60,000 ± 25,000
Yanovskiy et al. 5

Second, risks of other people are often wrongly estimated,
particularly when corporate employers make decisions for
their employees.
35
The authors of the latter paper
35
refer to
the well-known principal-agent problem, when agent may
act in a way that is contrary to the best interests of the
principal. Let us emphasize that the principal-agent problem
seems to be at least as severe between a citizen and a
governmental body.
Third and most important: let us address the question of
why from about 1995 to about 2015, VSL estimates by three
U.S. agencies—Department of Agriculture, Food and Drug
Administration and Environmental Protection Agency—
went up 3-fold from $2-4M to $9-10M (in 2019 dollars).
29
We should mention that any government agency estimating
VSL faces explicit conflicts of interest: the higher the VSL,
the better the outcome of cost-effectiveness analysis for any
proposed policy.
Consider, for example, a policy with a price tag $1000M
(one billion dollars). If VSL is $1.5 M, such a policy is
justified if it saves about 700 people; however, if VSL is
$10M—saving 100 people (that is, being 1/7 as effective)
is enough for a positive cost-effectiveness judgment. Prob-
ably not surprisingly, VSL estimations of up to US$70
million have been cited in the literature.
36
Because of high volatility of VSL values reported by dif-
ferent government agencies, and because of the explicit conflict
of interest in their VSL estimation, it seems to us prudent to use
our value for the cost-effectiveness threshold (CET).
As an example, let us apply cost-effectiveness analysis to
the COVID-19 crisis management in Israel. It can be esti-
mated that the direct economic cost of the lockdowns in
2020–2021 was about US$ 30 billion, while the Israeli
population is about 9.2 million, and GDP per capita is about
US$ 45,000 (see Supplemental Appendix). Dividing 30
billion by 1.4×45,000 yields about 500,000 quality-adjusted
life-years lost. The discussion of whether the above human
cost was justified is beyond the scope of this paper. We shall
just mention that the loss of 500,000 QALY in Israel is equal
to the loss of life due to cancer (about 11,000 deaths per
year,
37
12 life-years per death
38
) during 4 years.
Our analysis is not free from limitations. Several limita-
tions stem from our estimation of CET by analyzing salaries
in risky occupations.
The first limitation is that the data is more than four decades
old. Unfortunately, we could not find more up-to-date research
using this method which seems to us extremely important.
Next, one can reasonably ask whether individuals in jobs
with a substantive risk systematically over- or under-estimate
the risk. Really, personal perception can deviate from sci-
entific assessments. The known phenomena include some
underestimation of risks by young people and men as op-
posed to overestimation by elderly and women; overesti-
mation of low risks and underestimation of high.
39
However,
people usually perceive personal risks (including risk of
death) rather adequately, unlike risks to other people.
40
Another limitation is that hazard pay may be more so
received by individuals otherwise facing lower earnings
potential. If so, the wage rate needed to entice these indi-
viduals to face the added risks is likely to be much lower than
that required to entice the general population. The above
assumption, while plausible, is not necessarily correct. In
their estimation, Thaler and Rosen
25
took into account age,
gender, education etc. And simply speaking, the list of high-
risk occupations (Thaler and Rosen,
25
Table 1 at p. 288)
contains not a small number of highly-paid professions de-
manding well-educated employees—for example, electri-
cians, boilermakers, structural ironworkers, marshals,
constables, ships’officers, and even actors.
Further, we have not used discounting in translation from
VSL to CET per QALY. Proper way of discounting is an
unsettled question in health economics. For example, dif-
ferential discounting was used by in UK for some time,
41
with
1.5% per year for life and 6% for money. This practice,
however, has been terminated and substituted with equal
discounting at 3% per year.
41
Although discounting makes
estimations much more complex and prone to bias, we an-
ticipate that it will not affect the results in a serious manner. Let
us consider discounting in the three methods we used. Pro-
fessionals in risky occupations perform discounting subcon-
sciously and automatically, so another discounting would be
superfluous. The same can be said regarding CET estimation
based on Prospect Theory: people discount, subconsciously and
automatically, both their future earnings and their life-years
potentially lost. As for the legal practice, explicit discounting
of both life-years and earnings is actually performed,
42,43
so
once again, another discounting would be superfluous.
Additional limitation is connected to our use of Prospect
Theory. Certainly, life is very different from other goods. Not
only does the degree of importance people attach to the value
of life differ from person to person, but the relationship
between it and money may not be linear. However,
population-wide it seems that people do treat life like other
goods, which can be deduced, for example, from the analysis
of purchasing insurance.
28
We should also mention that in
health economics equal discounting is presently used for life
and money,
41
as discussed in the previous paragraph.
Last, our estimation was performed on the U.S. data only.
However, the practice of recommending CET values as a
fraction of GDP is well established and used, for example, by
WHO.
22
So, we believe that our estimation is generally
applicable to any country with comparable institutions.
Conclusions
Every public expenditure, including saving lives or extending
life expectancy of particular persons (target population), has
unwanted but unavoidable side effect of statistical shortening
of life expectancy of non-target population. Therefore, cost-
effectiveness considerations in healthcare and safety are not
technical, but ethical necessity. Population-wide, a life-
6INQUIRY

extending policy with cost-effectiveness below cost-
effectiveness threshold (CET) claims more life than it saves.
Because of high volatility of CET values reported by
different government agencies, and because of the explicit
conflict of interest in their estimations, it seems to us prudent
to use our value for CET, derived by three independent
methods and consistent with the current health policies:
CET = 1.0 ± 0.4 GDP (gross domestic product) per capita
per quality-adjusted life-year.
Therefore, 140% GDP per capita per quality-adjusted life-
year should be considered as the upper limit of prudent
expenditure on life extension.
Acknowledgments
The authors wish to thank Prof. Avi Caspi (Jerusalem College of Tech-
nology—JCT) for his encouragement of this work. We would like to thank
Prof. Shlomo Engelberg (JCT) and Prof. Eli Sloutskin (Bar Ilan Uni-
versity) for thorough reading the manuscript and suggesting many im-
portant improvements. We also wish to thank Dr Moti Brill (Nuclear
Research Center Negev, ret.), Prof. Noah Dana-Picard (JCT), the late Prof.
Ludwik Dobrzy´
nski (National Centre for Nuclear Research, Poland), Prof.
Marek Janiak (Military Institute of Hygiene and Epidemiology, Poland),
Dr Efraim Laor (Holon Institute of Technology), Prof. Michael Shapiro
(Technion –Israel Institute of Technology), Dr Barak Tavron (Noga Ltd),
and the late Prof. Alexander Vaiserman (Institute of Gerontology, Kiev,
Ukraine) for fruitful discussions and constructive criticism. Last but not
least, we would like to thank the anonymous reviewers whose constructive
criticism enabled to considerably improve the manuscript.
Author Contributions
MY and YS conceived the idea. MY, YYS and ONL performed the
ethical analysis. YS performed the data analysis. MY wrote the first
draft. MY, YS, YYS, ONL, AZ participated in finalizing the
manuscript. AZ performed the critical revision of the manuscript. All
the authors read and approved the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for
the research, authorship, and/or publication of this article: Jerusalem
College of Technology grant №5969.
ORCID iD
Moshe Yanovskiy, PhD https://orcid.org/0000-0002-7014-0522
Avi Zigdon https://orcid.org/0000-0003-1849-7206
Yehoshua Socol https://orcid.org/0000-0003-4167-248X
Supplemental Material
Supplemental material for this article is available online.
References
1. Sloan FA, Hsieh C-R. Health Economics. Boston MA: MIT
Press; 2017.
2. Stenberg K, Hanssen O, Edejer TT-T, et al. Financing transformative
health systems towards achievement of the health Sustainable
Development Goals: a model for projected resource needs in 67 low-
income and middle-income countries. Lancet Global Health. 2017;
5(9):e875-e887. doi:10.1016/S2214-109X(17)30263-2
3. Marseille E, Larson B, Kazi DS, Kahn JG, Rosen S Thresholds
for the cost-effectiveness of interventions: alternative ap-
proaches. Bull World Health Organ. 2014;93(2):118-124. doi:
10.2471/BLT.14.138206
4. Gold MR, Siegel JE, Russell LB, Weinstein MC, eds. Cost-
Effectiveness in Health and Medicine. New York: Oxford
University Press; 1996.
5. Viscusi WK, Aldy JE. The value of a statistical life: a critical
review of market estimates throughout the world. J Risk Un-
certain. 2003;27(1):5–76. doi: 10.1023/A:1025598106257
6. Neumann PJ, Sanders GD. Cost-effectiveness analysis 2.0.
N Engl J Med. 2017;376(3):203-205. doi: 10.1023/A:
1025598106257
7. Cameron D, Ubels J, Norstr ¨
om F. On what basis are medical
cost-effectiveness thresholds set? Clashing opinions and an
absence of data: a systematic review. Glob Health Action. 2018;
11(1):1447828. doi: 10.1080/16549716.2018.1447828
8. Bonis PAB, Wong JB A short primer on cost-effectiveness
analysis. UpToDate
®
. Sep 10, 2021. https://www.uptodate.
com/contents/a-short-primer-on-cost-effectiveness-analysis.
Accessed Dec 10, 2021.
9. HSE. Reducing Risk, Protecting People. HSE’s Decision-
Making Process. London: Health & Safety Executive; 2001.
10. United States Environmental Protection Agency, National
Center for Environmental Economics. Valuing Mortality Risk
Reductions for Environmental Policy: A White Paper; 2010.
https://www.epa.gov/sites/production/files/2017-08/documents/
ee-0563-1.pdf Accessed Dec 10, 2021.
11. Bosworth RC, Hunter A, Kibria A, The Value of a Statistical
Life: Economics and Politics. Logan UT: Strata; 2017. https://
strata.org/pdf/2017/vsl-full-report.pdf Accessed Jan 1, 2020.
12. Baker R, Chilton S, Donaldson C, et al. Searchers vs surveyors
in estimating the monetary value of a QALY: resolving a nasty
dilemma for NICE. Health Econ Pol Law. 2011;6(4):435-447.
doi: 10.101/51744133111000181
13. Nghiem S, Graves N, Barnett A, Haden C Cost-effectiveness of
national health insurance programs in high-income countries: A
systematic review. PLoS One. 2017;12(12):e0189173. doi: 10.
1371/journal.pone.0189173
14. Abraham J, Drake C, Sacks DW, Simon K Demand for health
insurance marketplace plans was highly elastic in 2014-2015.
Econ Lett. 2017;159:69-73. doi: 10.1016/j.econlet.2017.07.002
15. Blundell R, Stoker TM. Heterogeneity and aggregation. J Econ
Lit. 2005;43(2):347-391. doi: 10.1257/0022051054661486
16. Chernew M. Health care spending growth: can we avoid fiscal
armageddon? Inquiry: The Journal of Health Care
Yanovskiy et al. 7

Organization, Provision, and Financing 2010;47(4): 285-295.
doi: 10.5034/inquiryjrnl_47.04.285
17. Ehrlich I, Yin Y. Equilibrium health spending and population
aging in a model of endogenous growth: Will the GDP share of
health spending keep rising? J Hum Cap. 2013;7(4):411-447.
doi: 10.1086/675640
18. OECD. Health At a Glance 2019: OECD Indicators. Paris:
OECD Publication; 2019. doi: 10.1787/4dd50c09-en
19. Berwick DM. The moral determinants of health. JAMA. 2020;
324(3):225-226. doi: 10.1001/jama.2020.11129
20. Graham JD. Comparing Opportunities to Reduce Health Risks:
Toxin Control, Medicine and Injury Prevention. Dallas TX:
National Center for Policy Analysis; 1995.
21. Kip Viscusi V The value of life in legal contexts: survey
and critique. Am Law Econ Rev. 2000;2(1):195-210. doi:
10.1093/aler/2.1.195
22. Bertram MY, Lauer JA, De Joncheere K, et al. Cost-
effectiveness thresholds: pros and cons. Bull World Health
Organ. 2016;94(12):925-930. doi: 10.2471/BLT.15.164418
23. Woods B, Revill P, Sculpher M, Claxton K. Country-level cost-
effectiveness thresholds: Initial estimates and the need for
further research. Value Health. 2016;19(8):929-935. doi: 10.
1016/j.jval.2016.02.017
24. Shiroiwa T, Sung Y-K, Fukuda T, Lang H-C, Bae S-C, Tsutani
K. International survey on willingness-to-pay (WTP) for one
additional QALY gained: what is the threshold of cost effec-
tiveness? Health Econ. 2010;19(4):422-437. doi: 10.1002/hec.
1481
25. Thaler R, Rosen S. The value of saving a life: evidence from the
labor market. In: ET Nestor, ed. Household Production and
Consumption. Cambridge, MA: NBER; 1976:265-302.
26. Lawrence HO, Samuel HW. The Annual Consumer Price Index
for the United States, 1774 to Present, MeasuringWorth
website. Available at: http://www.measuringworth.com/uscpi/.
Updated 2021. Accessed Dec 10, 2021.
27. Kahneman D Thinking, Fast and Slow, USA: Farrar, Straus and
Giroux; 2011.
28. Tversky A, Kahneman D Advances in prospect theory: cu-
mulative representation of uncertainty. J Risk Uncertain. 1992;
5, 297–323. DOI: 10.1007/BF00122574
29. Merrill D. No One Values Your Life More than the Federal Gov-
ernment. Bloomberg; 2021. Available at: https://www.bloomberg.
com/graphics/2017-value-of-life/ Accessed Dec 10, 2021.
30. World Bank. GDP Per Capita –United States; 2021. Available at:
https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=
US Accessed Dec 10, 2021.
31. Marmot M. The Health Gap: The Challenge of an Unequal
World. London, England: Bloomsbury Publishing PLC; 2016.
32. Birch S, Gafni A. The biggest bang for the buck or bigger bucks
for the bang: the fallacy of the cost-effectiveness threshold.
J Health Serv Res Pol. 2006;11(1):46-51.
33. Padula WV, Chen H-H, Phelps CE Is the choice of cost-
effectiveness threshold in cost-utility analysis endogenous to
the resulting value of technology? a systematic review. Appl
Health Econ Health Pol. 2021;19:155–162. DOI: 10.1007/
s40258-020-00606-4
34. Appleby J, Devlin N, Parkin D. NICE’s cost effectiveness
threshold. BMJ. 2007;335(7616):358-359. doi: 10.1136/bmj.
39308.560069.BE
35. Chakravarty S, Harrison GW, Haruvy EE, Rutstr¨
om EE. Are
you risk averse over other people’s money? South Econ J. 2011;
77(4):901-913.
36. Viscusi WK. Mortality effects of regulatory costs and policy
evaluation criteria. Rand J Econ. 1994;25(1):94. doi: 10.2307/
2555855
37. Cancer Index. Israel Cancer Statistics. Available at: http://
www.cancerindex.org/Israel Updated; 2019. Accessed De-
cember 10, 2021
38. Song M, Hildesheim A, Shiels MS. Premature years of life lost
due to cancer in the United States in 2017. Cancer Epidemiol
Biomark Prev. 2020 December;29(12):2591-2598. doi: 10.
1158/1055-9965.EPI-20-0782. Epub 2020 Nov 13.
39. Andersson H, Lundborg P Perception of own death risk. J Risk
Uncertain. 2007;34(1):67-84.
40. Benjamin DK, Dougan WR, Buschena D. Individuals’esti-
mates of the risks of death: part II-new evidence. J Risk Un-
certain. 2001;22(1): 35-57.
41. Attema AE, Brouwer WBF, Claxton K. Discounting in eco-
nomic evaluations. Pharmacoeconomics. 2018;36:745–758.
See P.752, Table 1. doi: 10.1007/s40273-018-0672-z
42. Tinari FD, Kucsma KK. Assessing economic damages in
personal injury and wrongful death litigation: the state of New
Jersey. J Forensic Econ. 2010;21(2):219-234.
43. Taylor L, Brandt WG. Assessing economic damages in per-
sonal injury and wrongful death litigation: the state of Wash-
ington. J Forensic Econ. 2015;26(1):115-131.
8INQUIRY