Original Papers

Eur J Health Econom 2005 50:215222

DOI 10.1007/s10198-005-0294-1
Published online: 27. April 2005
Springer Medizin Verlag 2005

ealth expenditures are a significant

policy issue in many countries. The United States spends more on health care than
the entire gross domestic product of China. The health care expenditures in the
United States rose to an estimated $1.7 trillion in 2003, more than double the spending of $812 billion in 1992 and three times
that of $547.28 billion in 1987 [20]. Health
care spending varies significantly across
countries. In terms of purchasing power
parity Korea spent about US $809 per capita in 2000, compared to $4,887 spent by
the United States in 2001. As a proportion
of gross domestic product, Luxembourg
spent about 5 in 2001, compared 13.9
in the United States.
A variety of reasons have been adduced
to explain this rise in health expenditures.
Demographic change is one major cause. The proportion of the population aged
65 years or over varies from 5.7 in Turkey
to 18.1 in Italy. As health care advances,
curing diseases and prolonging life expectancy, more persons require health care. Since persons do not die easily, we have
a problem with the prevalence of disease
rather than its incidence. This dramatically changes the care from acute services to
rehabilitative ser vices. In addition, studies show that health care spending during
a persons lifetime is likely to be concentrated near death [2, 21]. Since elderly persons are at a greater risk of death, health
care spending is higher for them. Therefore the proportion of elderly in the population has a significant impact on overall
health care expenditures.

Vasanthakumar N. Bhat Hamden, CT, USA

Institutional arrangements
and efficiency of health care
delivery systems

Technology is another factor that drives

up the health care costs [14]. Many technologies are being adopted with only little
evaluation. Several technologies have been
found to have little or no effect in improving health outcomes [24]. New tests and
procedures have dramatically increased demand for health care services. Technology
has also increased the intensity and complexity of services requiring need for highly skilled employees. This increases wages and overall costs. In addition, information technology is dramatically changing
the way in which consumers interact with
the health care delivery system, leading to
a health care consumerism. Informed
consumers are demanding access to newer procedures and at the same time expecting better outcomes from their health care providers. The rising standard of living
is another factor that increases demand for
health services. According to Gerdtham
and Jonsson [8], increased per capita income has a positive impact on the per capita health care expenditures, and estimated income elasticity is higher than 0 and
approximately 1 or more than 1. The rising productivity and rising incomes have
changed expectations about health care.
A significant proportion of funding for
health care comes from taxes and social
security payments. In countries belonging
to Organization for Economic Co-operation and Development (OECD) the public proportion of health expenditure varied in 1998 from 44.8 in the United States to 92.4 in Luxembourg [19]. Waves of
health reform measures have been imple-

mented over recent years in every country.

The major goal of most reforms has been
to reduce costs. Without such reforms substantial increases in taxes and reductions
in benefits would have been necessary. Policy makers are thus faced with the daunting challenge of maximizing the value of
investments in health care.
The purpose of this study was to examine the effects of various health care financing, regulatory, and organizational designs
on the efficiency of health care delivery
systems (see [22]). We first discuss previous works, followed by the methods of data envelopment analysis (DEA) used to calculate the efficiency of health care delivery
systems and results of this analysis. Then
we examine the impact of various institutional arrangements on efficiency.

Previous works
Most analyses of international health care
delivery systems have focused on health
care expenditures. Such studies help to elucidate the impact of various policies and institutional arrangements; however, health
care expenditures are difficult to interpret
across countries as they are measured in
different currencies with vastly different
spending powers. In spite of such a drawback Newhouse [13] tried to identify factors that affect the quantity of health care
services in 13 different countries using 1971
data. Leu [11] extended the model using
1974 data from 19 OECD countries by adding more variables, including the proportion of the population under 15 and over
Eur J Health Econom 3 2005

| 215

Original Papers
Table 1

Health care resources and utilization, around 2000 (from [20])

Variable

Maximum

Minimum

Value

Country

Value

United
States
value

Practicing physicians
per 1000 populations

4.5

Greece

1.3

Turkey

2.7

Practicing nurses
per 1000 populations

14.7

Finland

1.7

Turkey

8.1

Acute care beds

per 1000 population

6.7

Luxembourg

2.2

Turkey

2.9

Long-term care beds per 1000

population aged 65+ years

76.3

Switzerland

3.8

Turkey

Magnetic resonance imaging

units per million population

23.2

Japan

2.0

Greece

Computed tomography

84.4

Japan

6.2

United
Kingdom

Consultations with
physicians per capita

14.4

Japan

2.5

Turkey

8.9

Diphteria-tetanus-pertussis
immunization rates for
young children

99.2

Sweden

82.1

United
States

82.1

Austria

74

Portugal

Hospital discharges
per 1000 population
Average length of stay
to acute care

281
9.6

Germany

3.8

Denmark

8.1
13.1

112
5.8

Inpatient surgical procedures 130

per 1,000 population

Austria

24

Turkey

82

Cardiac bypass procedures

per 100,000 population

United
States

17

Spain

205

205

65 years of age, public proportion of total

spending, and dummy variables for the
countries of United Kingdom, New Zealand, and Switzerland. Gerdtham et al. [8]
present a comprehensive analysis of health
care expenditures using data of 24 countries from 19701991 [16]. The goal of most
studies was to identify factors that affect
expenditures. However, our focus is on efficiency. Our goal is to examine efficiency in terms of inputs rather than expenditures on inputs. Efficiency is often defined
as the level of output that can be obtained
from a given mix of inputs. A simple efficiency measure is a ratio of aggregate
health outcome to aggregate health spending. However, health care delivery systems
involve multiple inputs and outputs. We estimate efficiency using DEA. Most applications of DEA in health care have been
in the evaluation of microlevel service efficiency [4, 9]. Only few studies have com-

216 |

Country

Eur J Health Econom 3 2005

pared efficiency across countries. Gupta et

al. [10] used DEA to calculate the efficiency of government spending on life expectancy and infant mortality, and Evans et al.
[6] to calculate the extent of health care attainment.

Methods
Per capita health expenditure is a poor indicator of the performance of a countrys
health care delivery system. Both in absolute terms and as a proportion of gross domestic product the United States spends
more on health care than any other country in the world and therefore could be assumed to have an inefficient health care delivery system. However, the United
States is more productive than many other nations if inputs are considered. . Table 1 presents the maximum and minimum values for health care resources and

utilization of 24 countries along with those for the United States. It is obvious that
the United States has fewer physicians,
nurses, and inpatient beds on a per capita
basis than several OECD countries with
lower per capita health care expenditures.
This demonstrates that performance measures based on expenditures are of little
use as different countries have different
prices for various inputs. The present used
DEA to calculate the efficiency of health
care delivery systems using health care inputs rather than their expenditures. The
concept of the efficiency of a health care delivery system represents its ability to
treat populations of various age groups
using health care resources such as physicians, nurses, inpatient beds, and pharmaceuticals. A country is said to be technically efficient if it cannot reduce its resource
use without some corresponding inability to treat some patients. Technical inefficiency can arises from a variety of sources
including the size of a country and an inappropriate supply of input resources. Analysts have traditionally estimated the efficiency of a health care delivery system
using simple ratios such as physicians per
population and nurses per bed. However,
a number of problems are associated with
such a simplistic multiple-ratio analysis. It
may be impossible to identify a country
which is most efficient in terms of all ratios. If a country does not excel in all ratios, it is difficult to determine the weights
that each ratio should be given to compare
efficiency across countries. It should be
possible to compare the performance of
one country against another so that countries can identify sources of any inefficiency. The method typically used to compare
efficiency levels is that of frontier efficiency. In this method one first identifies best
practice frontiers. The frontier represents
the best performance. The efficiency is calculated based on the performance of countries on the frontiers.
DEA is a nonparametric frontier method that uses the linear programming technique to identify the frontier countries
and construct a convex piecewise linear
surface, or frontier, over these countries.
DEA is based on Farells [7] idea to characterize the most efficient production frontier as the piecewise linear convex hull
composed of the most efficient countries

Abstract
in the space of multiple outputs and multiple inputs. Several authors have suggested
mathematical programming methods to
implement this idea (e.g., [1]), but it was
Charnes et al. [3] who coined the term
data envelopment analysis. A detailed description of the DEA method is presented
by Seinford and Thrall [12, 23]. DEA begins with the definition of the unit of assessment, which is typically called the decision making unit (DMU). In each DMU
various resources, called inputs, are converted into outcomes, called outputs. Using inputs and outputs of all DMUs we develop a production possibility set using
assumptions such as that interpolated input-output combinations are feasible, inefficient input-output combinations can exist, and output cannot be produced without any input. Production can be subject
to either constant returns to scale (CRS)
or variable returns to scale (VRS). In CRS,
when all inputs are increased by a certain
percentage, outputs also increase by the
same percentage. In VRS, on the other
hand, when all inputs are increased by a
certain percentage, outputs increase by either a lower or higher percentage. In other words, the VRS production exhibits either economies or diseconomies of scale. Efficiency is calculated by taking the ratio of the weighted sum of outputs to the
weighted sum of inputs. The weights are
chosen so as to maximize each DMU ratio. Two major advantages of DEA are that
it is not necessary to specify the form of
the production function or error distribution, and that it can handle more than one
output and input. However, DEA does not
have a probabilistic component, and therefore any deviation from the frontier is considered an inefficiency.
The goal of this study was to calculate the efficiency of health care delivery
systems of each selected country. To apply DEA we categorize each country as a
DMU. We classify the countrys population
into three age groups: 019, 2064, and
65 years of age or older. Such a classification is reasonable as health care spending
increases with age. Sur vey data indicate
that noninstitutionalized elderly persons
aged 65 years or over require an average of
six times the health care resources as persons under the age 18 years. Persons aged
65 years or over consume almost three ti-

mes as much health care as those aged 18

64 (Agency for Health Care Research and
Quality, Household Component Full Year
Files, 1996, Rockville, Md., USA). We assume that each country uses inputs, namely physicians, nurses, inpatient beds, and
pharmaceuticals to provide health care to
its residents. We assume that resources
required to satisfy the health care needs
of each group in the population are the
same across counties. This is not restrictive as the same assumption is also made in models involving multiple linear regression analysis. We use the numbers of
practicing physicians, number of practicing nurses, inpatient beds, and pharmaceutical consumptions as inputs and population aged 019 years, 2064, and 65 or
older as outputs. Volume of pharmaceuticals consumed are estimated in pharmaceutical purchasing power parity and expressed in United States dollars. The data for the year 1996 is from OECD [20] for
the year 1996. Statistical significance was
set at the level of P0.05.

Analysis of results
We examined 24 OECD countries in this
study. The profiles of these 24 countries
are given in . Table 2. We estimated the
efficiency of countries assuming CRS [5].
The CRS assumption reflects the fact that
output increases or decreases by the same
proportion as inputs, or example, that doubling all inputs doubles all outputs. Since
this is more likely to be true for a health
care delivery system, we use a CRS model. The efficiency under CRS are presented in . Table 3. Of the 24 countries 8 fall
on the frontier and have an efficiency of 1
under CRS. Denmark, Japan, The Netherlands, Norway, Portugal, Sweden, Turkey,
and the United Kingdom have an efficiency of 1. Belgium, Iceland, and Australia have the lowest CRS efficiency. One major
advantage of the DEA is its ability to identify peer countries with an efficiency of 1
that an inefficient country can compare itself against, for example, peer countries
for Australia are Sweden, Turkey, and the
United Kingdom. This implies that Australia consumes more inputs than a weighted average of inputs used by Sweden, Turkey, or the United Kingdom to produce
the same amount of outputs. Therefore if

Eur J Health Econom 2005 50:215222

DOI 10.1007/s10198-005-0294-1
Springer Medizin Verlag 2005

Vasanthakumar N. Bhat

Institutional arrangements
and efficiency of health care
delivery systems
Abstract
This study examined the efficiency of health
care delivery systems in 24 OECD countries.
Practicing physicians, practicing nurses, inpatient beds, and pharmaceuticals were considered as inputs to treat populations of various age groups. Data envelopment analysis
(DEA) was utilized to calculate efficiency. We
also calculated input efficiency that should
be helpful in determining excess number
of physicians, nurses, inpatient beds, and
pharmaceuticals consumed. Institutional
arrangements affect efficiency: public-contract and public-integrated countries are
more efficient than public-reimbursement
countries. Countries in which physicians are
paid in wages and salaries and countries
with capitation have higher efficiency than
fee-for-service countries. Countries in which
a primary care physician acts as a gatekeeper are also more efficient than countries
without gatekeepers.
Keywords
Delivery of health care OECD countries
Data envelopment analysis Health system
performance Health system efficiency

Eur J Health Econom 3 2005

| 217

Original Papers
Table 2

Profiles of countries (1996) (PPP purchasing power parity)

Health care
price levelsa

Total health expenditureb

Per capita
PPP (US $)

% of GDP

Public
expenditure (%)

Australia

0.703

1,874

8.4

66.1

Austria

0.938

1,862

8.3

69.7

Belgium

0.859

1,962

8.8

72.6

Canada

0.691

2,093

70.8

Denmark

1.072

2,003

8.3

82.4

Finland

1.015

1,486

7.6

75.8

France

0.846

1,987

9.5

76.1

Germany

0.857

2,340

10.9

76.8

Greece

0.528

1,275

9.6

53

Iceland

1.014

1,928

8.4

83.3

Ireland

0.823

1,239

6.6

71.4

Italy

0.834

1,566

7.5

71.8

Japan

1.018

1,686

6.9

78.4

Luxembourg

0.932

2,192

6.4

92.8

New Zealand

0.672

1,261

7.2

76.7

Norway

1.069

2,026

7.9

84.2

Portugal

0.737

1,195

8.5

64.7

Spain

0.761

1,217

7.6

72.4

Sweden

1.116

1,779

8.4

86.9

Switzerland

1.327

2,615

10.4

54.7

The Netherlands

0.642

1,818

8.3

66.2

Turkey

0.443

234

3.9

69.2

United Kingdom

0.854

1,440

82.9

United States

1.380

3,792

13.2

45.6

a From Purchasing Power Parities and Real Expenditures, OECD 2002, calculated from Tables 2

and Table J for the year 1999; b From [20]

Australia wants to improve its efficiency,

it can compare itself against Sweden, Turkey, and the United Kingdom. Peer countries for each country with an efficiency
less than 1 are given in . Table 4. Turkey
appears as a peer country for 14 countries
with efficiency under 1 and Sweden and
the United Kingdom for 10 countries with
efficiency under 1.
Noninstitutional factors such as age
structure, tobacco consumption, and alcohol consumption can affect efficiency.
We considered the age structure of populations in the DEA analysis. Tobacco use
contributes significantly to ill health. Studies show a high correlation between tobacco consumption (measured in grams per

218 |

Eur J Health Econom 3 2005

capita) and deaths from lung cancer, with

a 20-year time lag [19]. According to Or,
after controlling for various factors, increased smoking is associated with higher levels of premature mortality [18]. Since countries with a higher smoking rate are likely to require higher input resources, they should be less efficient. However, Spearmans correlation between the
proportion of self-reported daily smokers,
persons aged 15 years or over, and CRS
efficiency is not statistically significant
(r=0.2867, P=0.17, n=24). Alcohol consumption (in liters) per population aged
15 years or over also shows no statistically
significant association with CRS efficiency (r=0.1416, P=0.5092, n=24).

Public funding of health care includes

financing by central, state, or local governments, by social security schemes, and by
public investment in health facilities. It is
possible that countries with a larger proportion of public funding are likely to be
more efficient as government can impose
tighter controls on costs. However, Spearmans rank correlation between public
funding as a percentage of total health expenditure in 1996 and CRS efficiency indicates no statistically significant association (r=0.0695, P=0.7468, n=24).
Institutional arrangement can significantly affect efficiency. For the purpose
of examining health care costs, Hurst categorizes countries into public-reimbursement, public-contract, and public-integrated models [15] (. Table 4). The reimbursement approach involves payments
to providers retrospectively for ser vices
provided to patients. In the contract approach providers enter into a prospective agreement containing the terms and
conditions of payments for ser vices provided to patients. The contract approach
provides payers with better control over
spending than the reimbursement approach. In integrated health systems the
payers are responsible for both payments
and provisions for services. According to
Hurst, public-reimbursement countries
have the weakest and integrated countries
the strongest success in controlling costs.
One-way analysis of variance reveals that
the average efficiency of countries with
these approaches is significantly different.
The average efficiency in integrated and
public-contract countries is much higher
than that in public-reimbursement countries. According to the Tukey-Kramer multiple comparison test, differences in mean efficiency between public-reimbursement and integrated models and publicreimbursement and public-contract countries are statistically significant. However,
differences between efficiency of integrated and public-contract countries are not
statistically significant. Public-reimbursement countries have an average efficiency of 0.82 compared to 0.93 in integrated
model countries and 0.94 in public-contract countries. This is consistent with the
postulate of Hurst that public-reimbursement models have the lowest average efficiency.

Table 3

We also analyzed the impact of the

ways in which physicians are remunerated
in the ambulatory sector on CRS efficiency. Countries are categorized into fee-forservice, capitation, and wages and salaries. In fee-for-service countries a physician
is paid based on the services provided. In
capitation system a fixed amount per period is paid to a physician for each patient
registered with him. Fee-for-service countries have an average efficiency of 0.87, capitation countries of 0.90, and wages and
salary countries of 0.99. Although differences in these efficiency are not statistically significant, the mean of CRS efficiency
for fee-for-service countries is much lower than countries with capitation and wages and salaries. This indicates a need for
higher input resources in countries with
fee-for-service remuneration. This is consistent with the finding that in countries
in which remuneration of physicians is
on a fee-for-service bases, a higher number of physicians are likely to result in higher health expenditure after controlling for
other factors [16].
We also examined the affect on efficiency of an ambulatory gatekeeper system
in which a patient is required to have a referral from a general practitioner for nonemergency access to a specialist. The average of efficiency in countries with primary physicians as gatekeepers is 0.92, compared to 0.88 in countries without gatekeepers. Countries with primary physicians as gatekeepers have a higher average
efficiency than those without gatekeepers,
although the difference does not reach the
level of statistical significance. This is consistent with the finding that countries with
primary physicians as gatekeepers for inpatient care have lower overall expenditures
[16]. The average of input efficiency based
on various institutional arrangements is
presented in . Table 3.
Estimates of health care efficiency indicate how well a country utilizes its health
care resources. As a result health care efficiency may not be directly related to the
per capita health expenditures if prices of
inputs differ across countries. Although
the United States ranks 18th among the 24
countries in efficiency, it has the highest
per capita health care expenditures. There are several reasons for high costs. The
prices paid for input resources are the high-

Constant return to scale efficiency, rank, and input efficiency

CRS
Rank
efficiency
Country
Australia
Austria
Belgium
Canada
Denmark
Finland
France
Germany
Greece
Iceland
Ireland
Italy
Japan
Luxembourg
New Zealand
Norway
Portugal
Spain
Sweden
Switzerland
The Netherlands
Turkey
United Kingdom
United States

Input efficiency

0.75
0.98
0.70
0.87
1.00
0.93
0.81
0.84
0.98
0.74
0.87
0.81
1.00
0.78
0.87
1.00
1.00
0.999
1.00
0.86
1.00
1.00
1.00
0.83

22
11
24
15
1
12
19
17
10
23
14
20
1
21
13
1
1
9
1
16
1
1
1
18

1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.813
0.938
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.949
1.000
1.000
1.000
1.000
1.000
1.000

1.000
0.515
0.984
1.000
1.000
0.947
1.000
0.620
0.907
1.000
1.000
1.000
1.000
0.757
0.623
1.000
1.000
1.000
1.000
0.749
1.000
1.000
1.000
1.000

0.761
1.000
1.000
0.670
1.000
0.736
1.000
1.000
1.000
0.701
0.554
1.000
1.000
0.949
0.719
1.000
1.000
1.000
1.000
0.986
1.000
1.000
1.000
0.881

0.935
1.000
1.000
0.870
1.000
1.000
0.683
1.000
0.869
1.000
0.978
0.955
1.000
1.000
1.000
1.000
1.000
0.541
1.000
1.000
1.000
1.000
1.000
0.705

0.82

1.000

0.927

0.939

0.903

0.94
0.93

5
12

1.000
0.983

0.827
0.913

0.934
0.906

0.974
0.955

Physician
remuneration
0.87
Fee-for-service
0.90
Capitation
Wages and salaries 0.99

13
6
5

0.986
0.990
0.990

0.858
1.000
0.989

0.920
0.876
0.947

0.928
0.989
0.908

Primary physician
as a gatekeeper
Yes
No

13
11

0.991
0.878

0.904
0.983

0.896
0.940

0.950
0.937

Supply
arrangements
Publicreimbursement
Public-contract
Public-integrated

0.92
0.88

est in the United States (see . Table 2).

For example, the annual gross income of
a physician in the United States in 1996
was $199,000, compared to $56,818 in Australia and $69,412 in Finland [17]. The price of the health care here is expressed in
the United States dollars converted into local currency using exchange rates
that need to be paid in a country to buy

health care worth $1 in OECD countries.

The price of health care in the United States is almost 38 higher than the average
price in OECD countries. The utilization
rate of the resources is low in the United
States Although this helps to provide services without much waiting, it raises costs
significantly. For example, the occupancy
rate of inpatient beds in the United States
Eur J Health Econom 3 2005

| 219

Original Papers
Table 4

Efficient peers and institutional arrangement (1996)

Australia

Efficient peers

Supply arrangement

Remuneration

Primary physician
as gatekeeper

Sweden, Turkey, United Kingdom

Public- Reimbursement

Fee-for-service

No

Austria

Denmark, Norway, Sweden, Turkey

Public-contract

Fee-for-service

Yes

Belgium

Portugal, Turkey, Sweden, United Kingdom

Public-reimbursement

Fee-for-service

No

Canada

Sweden, Turkey, United Kingdom

Public-contract

Fee-for-service

Yes

Denmark

Public-integrated

Capitation

Yes

Finland

Netherlands, Norway, Turkey

Public-integrated

Wage and Salary

No

France

Portugal, Sweden, Turkey, United Kingdom

Public-reimbursement

Fee-for-service

No

Germany

Norway, Sweden, Turkey, United Kingdom

Public-contract

Fee-for-service

Yes

Greece

Portugal

Public-integrated

Fee-for-service

No

Iceland

Norway, Sweden, Turkey

Public-integrated

Capitation

Yes

Ireland

Sweden, Turkey, United Kingdom

Public-integrated

Capitation

Yes

Italy

Japan, Portugal, United Kingdom

Public-integrated

Capitation

Yes

Japan

Public-reimbursement

Fee-for-service

No

Luxembourg

Norway, United Kingdom, Turkey

Public-reimbursement

Fee-for-service

No

New Zealand

Norway, Turkey, United Kingdom

Public-integrated

Fee-for-service

Yes

Norway

Public-integrated

Fee-for-service

Yes

Portugal

Public-integrated

Wage and salary

Yes

Spain

Portugal, Sweden, Turkey

Public-integrated

Wage and salary

Yes

Sweden

Public-Integrated

Wage and salary

No

Switzerland

Netherlands, Norway, Turkey

Public-reimbursement

Fee-for-service

No

The Netherlands

Public-contract

Capitation

Yes

Turkey

Public-contract

Wage and salary

No

U nited Kingdom

Public-integrated

Capitation

Yes

United States

Sweden, Turkey, United Kingdom

Public/private-reimbursement

Fee-for-service

No

is about 65 compared to 88.7 in The

Netherlands and 84.3 in Japan [17].

Conclusions
This investigation used DEA to calculate
the efficiency of health care delivery systems in 24 OECD countries. We also present CRS and input efficiency. The efficiency was calculated with respect to the countries on the frontier. This does not mean that the countries on the frontier cannot increase their efficiency. The input efficiency should be useful for countries to
understand the extent of overuse of input
resources. Countries should also be able
to benchmark their performance against
their peers and identify policy measures
to improve efficiency. Institutional arrangements have impact on the efficiency. Public-contract and public-integrat-

220 |

Eur J Health Econom 3 2005

ed countries are more efficient than public-reimbursement countries. Countries in

which physicians are paid in wages and
salaries and countries with capitation have higher efficiency than fee-for-service
countries. Countries in which primary care physicians act as gatekeepers are also
found to be more efficient than countries
without gatekeepers.

Corresponding author
Vasanthakumar N. Bhat
Lubin School of Business, Pace University,
New York, NY 10038, USA
435 Waite Street, Hamden, CT 06517
e-mail: vbhat@pace.edu
Conflict of interest: No information supplied

References
1. Afriat SN (1972) Efficiency estimation of production functions. Int Econ Rev 13:568598
2. Brockman H (2002) Why is less money spent on
health care for the elderly than for the population?
Health care rationing in German hospitals. Soc Sci
Med 55:593608
3. Charnes A, Cooper WW, Rhodes E (1978) Measuring the efficiency of decision making units. Eur J
Operation Res 2:429444
4. Chirikos TN, Sear AM (2000) Measuring hospital efficiency: a comparison of two approaches. Health
Serv Res 34:13891408
5. Coelli TJ (1996) A guide to DEAP version 2.1: a data
envelopment analysis (computer) program. CEPA
working paper 96/8. University of New England:
Armidale
6. Evans DB, Tandon A, Murray CJM, Lauer JA (2000)
The comparative efficiency of national health systems in producing health: an analysis of 191 countries. GPE discussion paper series, no. 29, EIP/GPE/
EQC, World Health Organization
7. Farrell MJ (1957) The measurement of productive
efficiency. J R Stat Soc series A 120 part 3

8. Gerdtham UG, Jonsson B (2000) International comparisons of health expenditures: Theory, data and
econometric analysis. In: Culyer AJ, Newhouse
JP (eds) Handbook of health economics. Elsevier:
Amsterdam
9. Grosskopt S, Valdmanis V (1987) Measuring hospital performance: a nonparametric approach. J
Health Econ 6:89107
10. Gupta SK, Verhoeven HM (1997) The efficiency of
government expenditure: experiences from Africa. IMP working paper no. WP/97/15 International
Monetary Fund: Washington
11. Leu RE (1986) The public-private mix and international health care costs. In: Culyer AJ, Honsson B
(eds) Public private health services. Basil Blackwell:
Oxford
12. Lovell CAK (2001) Production frontiers, productive
efficiency. In: Fried HO, Lovell CAK, Martin AB, Whittle LS, Levit KR (eds) Trends in state health care expenditures and funding: 19801998. Healthcare
Financ Rev 22, http://www.meps.ahrq.gov/data_
pub/hc_fydata96.htm
13. Newhouse JP (1992) Medical care costs: How
much welfare loss? J Econ Perspect 6:321
14. Newhouse JP (1977) Medical care expenditure: a
cross-national survey. J Hum Resources 12:115
125
15. OECD (1992) The reform of health care: A comparative analysis of seven OECD countries. Health Policy Stud no. 2
16. OECD (1994) Health care reform: controlling spending and increasing efficiency. Economics Department working papers no. 149
17. OECD (2000) Health data 2000: a comparative analysis of 29 countries. Organisation for Economic Cooperation and Development: Paris
18. OECD (2000) Determinants of health outcomes in
industrialized countries: a pooled cross-country, time series analysis. OECD Econ Stud 30
19. OECD (2001) Health at a glance. Organisation for
Economic Co-operation and Development: Paris
20. OECD (2003) Health data 2003: a comparative analysis of 29 countries. Organisation for Economic Cooperation and Development: Paris
21. Roos NP, Montogomery P, Roos LL (1987) Health
care utilization in the years prior to death. Milbank
Q 65:231254
22. Schmidt SS (ed) The measurement of productive efficiency Oxford University Press, New York
23. Seinford ABLM, Thrall RM (1990) Recent developments in DEA: the mathematical approach to frontier analysis. J Econometrics 46:738
24. Weisbroad BA (1991) Health care quadrilemma:
an essay on technological change, insurance, quality of care and cost containment. J Econ Literature
29:523552

Eur J Health Econom 3 2005

| 221