Economic effects of echinococcosis
Department of Veterinary Microbiology and Parasitology, Faculty of Veterinary Medicine, University College Dublin, Shelbourne Road,
Ballsbridge, Dublin 4, Ireland
Cystic echinococcosis (CE) has a number of important economic effects. The most tangible of these is the cost of
expensive medical treatment for human cases. Each confirmed case of CE can cost the health services or individual
several thousand dollars. In addition to these costs, the additional cost of loss of edible offal from agricultural animals
is well known. This may result in the entire loss of an infected organ or at least the trimming and downgrading of that
organ, depending on local legislature. However, these losses may only be a relatively small percentage of the economic
losses attributed to CE. Recent evidence suggests, through quality of life surveys, that patients treated for CE never
fully recover and have a significant and permanent decreased quality of life. This has yet to be translated into monetary
terms, but it almost certainly will result in the loss of income, possibly through a lower paid job, and/or the additional
expense of increased ill health. Furthermore, in most reports, between 1 and 2% of CE cases are fatal. The death of
these individuals results in the loss of the potential lifetime’s economic output of these individuals. With alveolar
echinococcosis the mortality rate is much higher and such consequences more severe. There is also a considerable
amount of Soviet literature, and small amounts published elsewhere which suggests that CE also significantly affects
animal productivity. Thus, infected sheep tend to give birth to fewer lambs, have lower levels of food conversion,
produce less milk and have poorer quality fleeces then non-infected sheep. The total cost of the disease is the sum of the
various costs to the health services, costs of morbidity and losses in animal productivity. Due to the uncertainty of many
of these costs, it is appropriate to model these losses using techniques that can give a range of cost estimates. By using
analytical techniques such as Monte-Carlo analysis, on parameters that are difficult to determine accurately, all such
variables can be randomly varied simultaneously along likely frequency distributions. The results of this give a useful
sensitivity analysis of economic costs. In addition, the purchasing power of money in the local economy must also be
taken into account. One US $ buys much more in a developing country than in an industrialized economy.
Consequently, each lost $ will be more acutely felt in poor countries. Estimates of the financial burden of disease are
beneficial in deciding priorities for control. They are also potentially useful tools to lobby donors or non-governmental
organizations to fund control programs in poor countries.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Cystic echinococcosis; Economic effects; Monte-Carlo analysis
* Present address: Institute of Parasitology, Winterthurestrasse 266a CH 8057, Zu ¨rich, Switzerland. Tel.: ? /41-1-6358535; fax: ? /41-
E-mail address: email@example.com (P.R. Torgerson).
Acta Tropica 85 (2003) 113?/118
0001-706X/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
It is becoming increasingly important to study
the economic effects of echinococcosis, as such
effects are vital in the evaluation of potential
control programs. Some years ago the WHO
recommended that an economic evaluation of the
effects of parasitic zoonoses should be an integral
part of any control program (WHO, 1979). It is
important to gain an understanding of the poten-
tial financial implications of the disease as this can
be used to define cost-effective control programs.
Indeed such knowledge can be used to lobby for
funding from local governments, donor organiza-
tions or non-governmental bodies. Since cystic
echinococcosis (CE) is a disease affecting both
man and agricultural animals it has potentially
economic effects to both agriculture and to human
health. Thus it is important to develop a financial
approach to calculate the losses from both sectors
and to gain an overall insight into the cost of the
disease (Torgerson, 1997). It will also enable a
decision-making framework to prioritize control
of echinococcosis from scarce resources from both
the agricultural and health budgets. This paper
aims to summarize the possible economic effects of
echinococcosis and examine methodology to get
an understanding of how these economic effects
can be calculated. This includes mathematical
techniques to overcome some of the problems
associated with uncertain estimates of some of
the cost variables. Finally, these ideas and techni-
ques will be illustrated by applying them to
Jordan, a developing country where CE is a
significant problem. This paper will only examine
the potential costs of the disease and not the costs
or financial benefits of a control intervention.
2. Human health costs
The most obvious human health cost and indeed
the easiest to define is the cost of medical and or
surgical treatment of patients suffering from
clinical echinococcosis. This cost is relatively easy
to calculate by obtaining a representative sample
of case records and costing the range of interven-
tions those patients have. There will be some
variability between individual cases and in the
method by which it is costed. In some health care
systems, where the cost of treatment is essentially
private (even when the population has medical
insurance support); hospitals and physicians will
have charges for each intervention. In a publicly
funded health service, such as the UK, the costs
may not be so obvious. However, even in the UK,
there are internal auditing procedures that allow
the costs of individual procedures to be calculated
(Torgerson and Dowling, 2001).
There are a number of other potential effects on
human health that are not so readily defined. This
takes the form of human morbidity and mortality
costs. The question of the long term health of an
individual who has either been treated for echino-
coccosis or is unwittingly affected by undiagnosed
disease should be considered. In a recent paper
Torgerson and Dowling (2001) demonstrated that
the long term quality of life in-patient treated for
CE was permanently affected. Although no finan-
cial cost was calculated, it has to be assumed that
such long term quality of life impairment must
have some effect. This could be due to individuals
having impairment in the workplace and thus
having less highly paid work than would otherwise
have been the case, having greater levels of
absenteeism from work or possibly requiring
greater levels of long term nursing care. It may
be possible that individuals with undiagnosed
echinococooisis may also have a decreased quality
of life, although that study has not yet been
Finally, the cost of fatal cases of echinococcosis
must be considered. The mortality rate for CE is
generally reported at 1?/2%, although this is much
higher with AE. The value of human life is still a
controversial subject and has been calculated in a
number of ways. The capital approach calculates
the potential loss of income form the time of
death. This method has a number of advantages;
most notably that it is relatively easy to calculate.
Other methods include the willingness to pay
approach. Here the value of an individual’s life is
calculated based on the extra investment an
individual is willing to spend in order to reduce
the risk of premature death. Many health econo-
mists suggest that no value can be put on some-
P.R. Torgerson / Acta Tropica 85 (2003) 113?/118
one’s life or even their health. In this case,
concepts known as quality of life years or dis-
ability-adjusted life years are often used (McGuire
et al., 1988; World Bank, 1993). Whilst this has
advantages in comparing different strategies in
terms of quality of life years saved, because CE has
significant economic effects on livestock a finan-
cial approach is more appropriate.
With all these human health costs the totals are
normally less in poor developing countries than
rich industrialized countries. This is because hos-
pital charges tend to be smaller because medical
services tend to be less sophisticated and less well
equipped in poorer countries and because labor
charges are lower. Likewise, any morbidity or
mortality costs will also be lower because labor
costs are cheap in poor countries.
3. Animal health costs
This is the other side to the cost associated with
CE. Again some costs can be precisely defined.
The main example is the costs of loss of edible
offal, mainly liver, from animals with CE. Depend-
ing on local legislature, such organs may be
completely condemned or diseased parts of the
organ trimmed. At worst such condemnations will
result in the total loss of that organ, although in
some countries such as Uruguay (Torgerson et al.,
2000), there is some possibility that condemned
organs may still have a value such as to the
However, potentially the largest costs to animal
health could be due to production losses in
livestock affected by CE. This can take the form
of a reduction in live weight gain, reduced yield of
milk, reductions in the fertility rates and reduc-
tions in the value of wool or other products. With
some of these losses estimated as 10% or more,
these production losses represent the most serious
CE-attributable losses to agriculture. Many of
these losses can be difficult to estimate as there
have been few controlled studies. Nevertheless,
data, such as from the former USSR (Kenzhebaev,
1985; Ramazanov, 1982) and data reviewed by
Polydorou (1981) do suggest that these losses are
potentially very significant and should not be
4. Model of the economic losses
To try to get an insight into to the economic
losses in a particular region or country a spread-
sheet model was constructed (Torgerson et al.,
2001). This model summed all the potential losses
due to the disease. Precisely known losses, such as
the cost of surgical treatment is based on the
numbers of people treated plus an estimate of the
per person treatment cost. Similarly, the loss of
edible offal is based on estimates of the prevalence
and the numbers of animals slaughtered per year
(adjusted for the average age of slaughter). The
less precisely known variables such as production
losses are estimated from reports from the litera-
ture. Some possible losses, such as long term
human morbidity, are not known at all. In these
categories a conservative estimated (i.e. low esti-
mate) are made of these losses.
To model uncertainty, in particular for the less
precisely known variables, Monte-Carlo techni-
ques have been written into the model. This
technique has also be used by Majorowski et al.
(2001) to estimate economic losses attributed to
CE in Tunisia. This allows for these variables to be
randomly varied along a likely distribution range
of the variable for each simulation undertaken.
With precisely known variables, such as hospital
costs, these are given a narrow range and a
distribution consistent with the observed data.
Hospital costs tend to fit a lognormal distribution,
but are a snapshot sample. The true mean hospital
costs should be within two standard errors of the
snapshot costs, thus this variable is given a
lognormal distribution with a standard deviation
equal to the standard error of the sample costs.
Variables known with less precision have to be
given a much wider confidence interval along
which to vary each simulation. Long term human
morbidity costs, for example, were given a gamma
distribution to ensure that on most simulations the
costs were very conservative (on the low side), but
also allowed for occasional very high costs which
may become apparent once further knowledge
P.R. Torgerson / Acta Tropica 85 (2003) 113?/118
becomes available and represents a worse case
scenario. Finally some variables were linked. For
example the animal population is inversely linked
to the market value of meat or milk products.
Thus, if on any one simulation, a higher than
average animal population is assigned, then a
lower than average price of meat products is
assigned. This is to model supply and demand in
the economy. The model is then run 10000 times
with each variable randomized for each simula-
This randomization approach has merits. First
it can give sensitivity analysis. This will give a
likely median scenario with 95% confidence limits
and a best and worst case scenario. Although some
of the variables are not well known, because they
are all varied simultaneously and randomly, the
laws of probability make it very unlikely that on
any one simulation all the variables will be chosen
to be either highly conservative or extreme values.
Thus a conservative value assigned to one variable
will be minimized by an extreme value of another
in the same simulation. This in itself is a useful
model of these unknown variables as, with further
research, it is unlikely that all imprecisely known
variables will be at the conservative range of the
estimates. This is illustrated by the coefficient of
variability. The results when the whole model is
run has a much lower coefficient of variability then
many of the individual parts of the model.
All long term losses need to be discounted for
the estimated period of that loss. This is because $1
at present day values could be invested and be
worth more in the future. With possible income
losses due to long term morbidity or death for
example, the present income has to be discounted
to the future for each year of the potential loss.
The World Bank (1993) recommended a discount
rate of 4%. Thus a premature death must be
discounted at a rate of 4% from the time of death
until the expected retirement age to obtain an idea
of the loss of income due to this fatality. However,
the rate of 4% may not always be appropriate. In a
rapidly growing economy, a rate greater than this
may be necessary.
6. The example of Jordan
Jordan is a lower middle income developing
country in the Middle East, which is highly
endemic for echinococcosis. Although levels of
infection in sheep approach 50% (Torgerson et al.,
1998), the annual human incidence rate is approxi-
mately 2.9 surgical cases per 100000 (Kamhawi,
1995). General socio-economic data for Jordan
and the estimated numbers of surgical cases are
illustrated in Table 1. An economic evaluation of
echinococcosis was undertaken in this country (a
more detailed analysis can be seen in Torgerson et
al., 2001) using the methods described above.
This resulted in a median estimate for the total
cost of hospital treatments to be $70500 with a
likely range (95% confidence limits) of $56071?/
87118. The other contributions to human health
losses are illustrated in Table 2, with the total
annual human health losses to be in the region of
$185000 (range $79571?/498011). However, this
loss was small compared with the possible pro-
ductivity losses in livestock (Table 3). The loss of
edible liver is estimated at over $850000 whilst
total losses to the livestock industry were over $3.5
million annually. This high figure for livestock
losses represents the potential impact that this
disease has on the agricultural industry. The total
Data used to calculate losses due to CE in Jordan
Annual surgical incidence
Estimated numbers of cases per year
Average age of treatment
Mean cost of treatment (range of 120
Prevalence in sheep
Prevalence in goat
2.9 per 100000
P.R. Torgerson / Acta Tropica 85 (2003) 113?/118
losses to the Jordanian economy is a median of
$3.9 million (95% confidence limits of $2.6?/6.5
million). Furthermore, because of the low purchas-
ing power parity in Jordan, these losses would be
the equivalent of $17 million in the USA.
7. Sensitivity analysis
Sensitivity analysis is a useful tool to model
worst and best cast scenarios for cost implications.
This has been widely used for other diseases (for
example Henderson et al., 1984). This techniques
was first suggested by Torgerson et al. (2000) for
the economic evaluation of echinococcosis and has
been further refined using the mathematical tech-
niques described above. With accurate data un-
available for some of the potential cost variables,
this approach gives a range of scenarios based on
different assumptions. In particular with conser-
vative assumptions it gives the lowest cost esti-
mates and if these turn out to be larger than the
maximum cost of control then it suggests that a
control program will always be financially viable.
Likewise if the maximum cost of the disease is less
than the minimum cost of an intervention program
then control is not economically viable. Intermedi-
ate situations may indicate that further research
needs to be undertaken.
Henderson, J.B., Beattie, C.P., Hale, E.G., Wright, T., 1984.
The evaluation of new services: possibilities for preventing
congenital toxoplasmosis. International Journal for Epide-
miology 13, 65?/72.
Kamhawi, S., 1995. A retrospective study of human cystic
echinococcosis in Jordan. Annals of Tropical Medicine and
Parasitology 89, 409?/414.
Kenzhebaev, S.A., 1985. Pokazateli eckonomicheskovo ush-
chererba pri eckhinokokkoze karakulskikh ovets. Trudy
Vsesoyuznogo Instituta Gel, mintologii-im K.L. Skryabina
28, 62?/66 (in Russian).
Majorowski, H., Carabin, M., Kilani, M., Bensalah, A., 2001.
Echinococcosis in Tunisia: an economic analysis. In:
Menzies, F.D., Reid, S.W.J. (Eds.), Proceedings of the
Society for Veterinary Epidemiology and Preventative
Medicine, pp. 187?/201.
McGuire, A., Henderson, J., Mooney, G., 1988. The Econom-
ics of Health Care: An Introductory Text. Routledge and
Kegan Paul, London.
Polydorou, K., 1981. Animal health and economics. Case
study: echinococcosis with a reference to Cyprus. Bulletin,
Office International des Epizooties 93, 981?/992.
Ramazanov, V.T., 1982. Evaluation of economic losses due to
echinococcosis. In: Lysenko, A. (Ed.), Zoonoses Control:
Collection of Teaching Aids for International Training
Course, vol. 2. Center of International Projects GKNT,
Moscow, pp. 283?/285.
Torgerson, P.R., 1997. The economic implications of zoonoses.
In: Holland, C.V. (Ed.), Modern Perspectives in Zoonoses.
Royal Irish Academy, Dublin, pp. 133?/144.
Torgerson, P.R., Dowling, P.M., 2001. Estimating the eco-
nomic effects of cystic echinococcosis. Part 2: an endemic
region in the United Kingdom, a wealthy industrialized
Human health costs in Jordan
Reduced quality of life
Morbidity due to undiagnosed
Loss of income in fatal cases
Total human health costs
Animal health costs in Jordan
Ninety five percent
confidence limits ($)
Loss of edible liver
Loss of lamb production
Loss of kid production
Loss of sheep meat
Loss of goat meat production
Loss of sheep’s milk
Loss of goat’s milk
Loss of value of fleece from
Total animal health losses
P.R. Torgerson / Acta Tropica 85 (2003) 113?/118
economy. Annals of Tropical Medicine and Parasitology 95, Download full-text
Torgerson, P.R., Williams, D., Abo-Shehada, M.N., 1998.
Modelling the prevalence of Echinococcus and Taenia
species in small ruminants of different ages in northern
Jordan. Veterinary Parasitology 79, 35?/51.
Torgerson, P.R., Carmona, C., Bonifacino, R., 2000. Estimat-
ing the economic effects of cystic echinococcosis: Uruguay,
a developing country with upper-middle income. Annals of
Tropical Medicine and Parasitology 94, 703?/713.
Torgerson, P.R., Dowling, P.M., Abo-Shehada, M.N., 2001.
Estimating the economic effects of echinococcosis. Part 3:
Jordan, a developing country of lower middle income.
Annals of Tropical Medicine and Parasitology 95, 595?/603.
WHO, 1979. Parasitic Zoonoses. Report of a WHO Expert
Committee with the Participation of FAO. Technical
Report Series 637. World Health Organization, Geneva.
World Bank, 1993. World Development Report Investing in
Health. Oxford University Press, Oxford.
P.R. Torgerson / Acta Tropica 85 (2003) 113?/118