Health Policy 95 (2010) 137–143
Contents lists available at ScienceDirect
Health Policy
journa l homepage: www.e lsev ier .com/ locate /hea l thpol
Disinve s to
healthc
Sabina N
Laboratorio Ma 33, 5612
a r t i c l
Keywords:
Health prioriti
Benchmarking
Efficiency
Health resourc
g servi
e aim
ly on h
07 data
impact on the level of resources used. For each indicator, the first step was to estimate the
gap between the performance of each Health Authority (HA) and the best performance or
the regional average. The second step was to measure this gap in terms of financial value.
The results of the analysis demonstrated that, at the regional level, 2–7% of the healthcare
budget can be re-allocated if all the institutions achieve the regional average or the best
1. Introdu
Resourc
require hea
constrained
ethical prin
ness” throu
argue for a
opposed to
The typ
gets based o
lead to sub-
approaches
tematic and
[2,3] even i
∗ Correspon
E-mail add
0168-8510/$ –
doi:10.1016/j.practice.
The implicationsof this studycanbeuseful forpolicymakersand theHAtopmanagement.
In the context of resource scarcity, it allows managers to identify the areas where the
institutions can achieve a higher level of efficiency without negative effects on quality of
care and instead re-allocate resources toward services with more value for patients.
© 2009 Elsevier Ireland Ltd. All rights reserved.
ction
e scarcity and increasing demand for services
lth systems to cope with difficult choices within
budgets. A range of concerns, ranging from
ciples such as “accountability for reasonable-
gh to economic goals of increasing productivity
thoughtful approach that targets reductions as
across-the-board cuts.
ical health system approach of deriving bud-
n historical spending or political pressures can
optimal use of limited resources [1]. Economic
can help decision makers by providing a sys-
explicit way to set evidence-based priorities
f they are not the sole consideration [4,5].
ding author. Tel.: +39 050883871; fax: +39 050883890.
ress: snuti@sssup.it (S. Nuti).
In the process of resource re-allocation, different
countries have followed varying approaches for setting pri-
orities at national level [6]. Since 1970s many countries
have adopted the Program budgeting and marginal analysis
(PBMA) in thehealth sector [4,7]. PBMAhasbeendeveloped
as an attempt to rationalize the incremental budgeting
approach, based on applications of opportunity cost and
marginal analysis [8]. PBMA can be deployed at the micro-
level (i.e. specific service areas or treatments) but also at
the meso-level (Health Authorities) and the macro-level
(Regional Health Systems or National Health Systems) [9].
Other budgeting and re-allocation techniques have used
Health Technology Assessment techniques to guide disin-
vestment decisions in ineffective treatments (e.g. guidance
on disinvestment from NICE) [10,11].
This paper describes a study carried out in the Regional
Health System of Tuscany, Italy. Using 2007 performance
data, the study measures the impact that performance
improvement could have on the amount of resources that
see front matter © 2009 Elsevier Ireland Ltd. All rights reserved.
healthpol.2009.11.011stment for re-allocation: A proces
are
uti ∗, Milena Vainieri, Anna Bonini
nagement e Sanità, Scuola Superiore Sant’Anna, piazza Martiri delle Libertà
e i n f o
es
es
a b s t r a c t
Resource scarcity and increasin
within constrained budgets. Th
the Tuscan Health System in Ita
re-allocation.
The analysis was based on 20identify priorities in
7 Pisa, Italy
ce demand lead health systems to cope with choices
of the paper is to describe the study carried out in
ow to set priorities in the disinvestment process for
benchmarking of the Tuscan Health System with an
138 S. Nuti et al. / Health Policy 95 (2010) 137–143
Health Authorities (HAs) and the Regional Health Sys-
tem could save and re-allocate to other services. This
exercise highlights the role that benchmarking best prac-
tices can play in disinvestment decisions. For this study,
“disinvestm
withdrawin
practices, p
that are de
tive to their
resources [
service red
ings achiev
benchmark
Perform
method in h
the debate
vice on how
models for
highly desi
an OECD re
improving p
ciency and t
rising dema
ing. In this s
disinvestm
by improvi
resource ut
2. Backgro
The use
be very us
is the wid
[14]. This is
(HAs) that h
used inter-
efforts. Du
gets, Health
reduce thei
and existin
Officers (CE
rion is not
rates are o
cases. They
have achiev
approach to
is no ability
basis.
This situ
the Region
mance Eval
12 Local He
tals (THs) [
up of 130 in
lation healt
satisfaction
performanc
to provide
opportunity
The maj
on the basis
dards available. When there is no standard the evaluation
is given on the basis of the regional mean or median. The
performance assessment is divided into five classes:
ry goo
od per
erage
or per
ry poo
asses
or th
ch ind
uring t
propo
in oth
itted a
eview
al con
gers, p
s the
s now
vels a
order
tifying
e was
s all
ted h
s that
the sa
2007 a
ercent
rmanc
poor p
with
07 and
ty and
teratu
gth of
on rat
Likew
lower
ethod
line
ts [13
rform
s, the r
on 20
rces t
nstru
entifi
tion o
can be
he calc
HA an
lation
osed o
alth a
tion anent” in healthcare describes the processes of
g health resources from existing healthcare
rocedures, technologies or pharmaceuticals
emed to deliver little or no health gain rela-
cost, and thus, no efficient allocation or health
10]. This means that disinvestment includes
uctions due to inappropriateness and sav-
ed through better efficiency identified through
ing (e.g. lower cost for the same output).
ance benchmarking is a common improvement
ospitals in Anglo-American countries. Despite
that has developed in UK National Health Ser-
best practices were disseminated and used as
emulation, benchmarking is still considered a
rable policy instrument [12]. Johnston [13] in
port suggests that benchmarking is useful for
erformance, particularly improvements in effi-
hat it may provide a valuable way of reconciling
nds for healthcare with limits on public financ-
ense benchmarking could be useful as a guide to
ent because it identifies where to free resources
ng performance indicators with an impact on
ilization.
und
of benchmarking as a managerial tool may
eful in the Italian health sector where there
espread belief that costs cannot be reduced
especially true within the Health Authorities
ave typically not competed with each other or
regional benchmarking to guide improvement
ring the evaluation of annual Regional bud-
Authorities typically argue that they cannot
r spending due to already constrained budgets
g deficits. In fact, most of HAs Chief Executive
Os) argue that the current capitation crite-
linked to true population needs and that DRG
ut-of-date, particularly for highly complexity
typically argue that it is self-evident that they
ed maximum efficiency. Without a systematic
evaluating the performance of each HA, there
to respond to these claims from an evidence
ation has changed in the Tuscany Region, where
has introduced a multidimensional Perfor-
uation System (PES) to assess and monitor its
alth Authorities (LHAs) and 4 Teaching Hospi-
15,16]. The PES is based on 50 measures, made
dicators, organized into six dimensions: popu-
h, regional policy targets, quality of care, patient
, staff satisfaction, and efficiency and financial
e. Each measure is benchmarked and published
management, providers, and consumers the
to compare the results across all organizations.
ority part of indicators receives an evaluation
of the international, national or regional stan-
1. ve
2. go
3. av
4. po
5. ve
Cl
mean
for ea
D
agers
used
subm
also r
annu
mana
guide
PES i
HA le
In
quan
manc
acros
repor
ment
were
In
the p
perfo
very
lated
in 20
quali
the li
in len
missi
[20].
with
3. M
In
repor
of pe
ence
based
resou
Co
the id
tifica
that
(3) t
each
trans
comp
of He
strucd performance;
formance;
performance;
formance;
r performance.
are identified on the basis of the standard, the
e median and the regional standard deviation
icator.
he pilot phase of PES, regional and HA top man-
sed a list of indicators, some of them already
er countries [17–19]. Clinical indicators were
nd approved by the Region. All indicators were
ed a variety of consensus exercises including an
sensus conferences involving HA and regional
rofessionals and consumers. The same process
inclusion of new indicators each year [15]. The
a central part of governance at the regional and
nd is linked to CEOs’ compensation.
to support regional and HAs management in
the achievable level of efficiency if perfor-
at the highest observed (benchmark) level
HAs, a research team presented the results
ere, that translate the performance improve-
the system could achieve if efficiency levels
me as benchmark performers.
nd 2008 the overall performance (calculated as
age of PES indicators with good and very good
e minus the number of indicators with poor and
erformance) is significantly and inversely corre-
the adjusted per-capita cost (r = −0.70, p < 0.05
r = −0.58, p < 0.05 in 2008). This finding, that
cost are inversely related, is consistent with
re for some indicators. For example, reductions
stay have not led to increases in 30-day read-
es or the volume of physician visits for patients
ise, shorter hospital stays have been associated
post-discharge death rates [21].
ology
with other European healthcare performance
], the Tuscan PES shows substantial variability
ance across HAs [22]. Based on these differ-
esearch team created an exploratory simulation
07 performance data to quantify the amount of
hat could be saved.
ction of the simulation followed four steps: (1)
cation of indicators to be analyzed, (2) the iden-
f the minimum and maximum improvement
required to achieve benchmark performance,
ulation of gaps between the performance of
d the mean or best performance, and (4) the
of gaps into financial terms. An advisory panel
f top managers from the Regional Department
nd HAs provided advice throughout the con-
d analysis of the simulation.
S. Nuti et al. / Health Policy 95 (2010) 137–143 139
Table 1
List of selected indicators.
No. Indicators Intervention
regarding
1 General practitioners expenses Primary care
2 Paediatrics expenses
3 Other services expenses
4 Pharmaceutical expenses Pharmaceuticals
5 Inappropriate hospitalization rate Hospital
6 Ave
DRG
7 Pre
8 Rea
9 % H
day
10 Pot
11 Abs
The firs
board valid
research tea
indicator se
Indicato
• Expected
to concre
literature
ommenda
• Indicators
latable in
The 11 i
primary car
resources m
cators, six a
(ALOS) for e
interventio
ceutical exp
focus on the
on these ind
For hum
potential re
number of
units/wards
Once this in
provides fo
or re-alloca
The abs
organizatio
employees
(i.e. the pe
there is a
organize se
The hosp
vides an est
1 The indica
than 60 and 65
when they are
extent the ind
builds off of work on both the use of small area varia-
tions in utilization management [24,25] and ambulatory
care sensitive conditions [26]. Reductions in inappropriate
or avoidable hospitalizations should reduce costs and pro-
vide opport
than 30 day
high value
organized a
considered
e read
is an
hat pr
ing sh
e seco
um a
pected
pothes
ained
ched t
pothes
ained
ional b
ble 2
he mi
e hos
an ho
durin
se an
as an i
ator th
e appr
t equiv
d ALO
ent au
8,29].
e thir
real p
o hyp
een th
djuste
e fina
iffere
od for
riticalrage length of stay for medical
s
-surgical length of stay
dmission within 30 days
ospitalization with LOS > 30
s
Continuity
ential retirees Human resources
management
enteeism rate
t step was to select indicators. The advisory
ated the criteria of selection put forward by the
m and selected 11 indicators from the 130 PES
t.
r selection was based on the following criteria:
improvement in the indicator had to be linked
te and feasible actions through evidence in the
or the consensus process or advisory board rec-
tions.
had to be expressed in Euros or easily trans-
to Euros.
ndicators described performance in hospitals,
e, pharmaceuticals, or on continuity and human
anagement (see Table 1). Among the 11 indi-
re efficiency indicators: average length of stay
ach medical DRGs, pre-surgical LOS for planned
ns, three primary care indicators and pharma-
enses (mostly linked to off-label use). They all
way resources have been used. Improvements
icators directly translate into potential savings.
an resources management, the indicator on
tirees1 shows the opportunities (in terms of
employees) to re-allocate resources among
/services by reducing the number of personnel.
dicator has been translated in financial terms, it
rward-looking statements of potential savings
Th
tions
HA t
plann
Th
minim
be ex
• Hy
obt
rea
• Hy
obt
reg
Ta
ses: t
for th
medi
mark
becau
seen
indic
vativ
effor
aroun
differ
[21,2
Th
HA’s
the tw
betw
HAs a
Th
cial d
meth
and ctions.
enteeism rate is an indirect indicator of the
nal climate. High rate of absence means that
are not satisfied with their working conditions
rception of equitable treatment) [23]. When
high rate of absence, HAs are obliged to re-
rvices with resulting extraordinary expenses.
italization rate over the regional median, pro-
imate of potential inappropriate care. This work
tor counts the number of employees with an age of more
. According to the Italian labour law clinicians could retiree
60, usually they end of working when they are 65. To this
icator highlights the number of potential retirees.
selected we
the indicat
different m
Table 3):
• The avera
retirees).
• The total
absenteei
• The total
ization ra
• The total
(ALOS andunities for disinvestment. Hospital stays longer
s provide an indicator of continuity of care: a
is a signal that primary and hospital care is not
round patients. These long lengths of stay were
partially avoidable.
mission rate within 30 days for similar condi-
indicator of preventable hospitalizations [27]:
ovided appropriate treatment and discharge
ould have a lower rate of readmission.
nd step was to point out, for each indicator, the
nd the maximum level of improvement that can
from HAs. The two hypotheses were:
is 1 (minimum improvement): changes to be
if all HAs with a poor and very poor performance
he mean or minimum standard.
is 2 (maximum improvement): changes to be
if all HAs could reach the regional target or
enchmark.
shows the two exceptions to above hypothe-
nimum and maximum improvement are equal
pitalization rate and the ALOS indicators. The
spitalization rate, was chosen as the bench-
g the consensus indicator selection exercise
hospitalization rate below the median could be
ndicator of poor access. Likewise, for the ALOS
e consensus exercise suggested that a conser-
oach keeping the benchmark and the minimum
alent because of limits on case mix adjustment
S. This is a conservative approach given that
thors have used more aggressive approaches
d step was to estimate the gap between each
erformance and the performance required by
otheses. This was done by taking the difference
e identified target for improvement and each
d observed performance from 2007.
l step was to determine the potential finan-
nce under each hypothesis. The appropriate
translating measures into costs has been widely
ly discussed [30–34]. One-third of indicators
re already valued in financial terms but where
ors were not valued in financial terms, three
ethods to attribute costs were applied (see
ge wage that would be eliminated (potential
personnel costs that would be avoided (the
sm rate).
excess volume of DRGs (inappropriate hospital-
tes).
direct costs linked to the excess use of beds
readmission rates).
140 S. Nuti et al. / Health Policy 95 (2010) 137–143
Table 2
The criteria applied for each indicators in the two hypotheses.
No. Interventions Hypothesis 1 Hypothesis 2
1 General practitioners expenses Regional mean Regional best performance
2 Paediatrics expenses
3 Other services expenses
4 Pharmaceutical expenses
5 Inappropriate hospitalization rate Regional med
6 Average length of stay for medical DRGs Regional mea
7 Pre-surgical length of stay Regional mea
8 Readmission within 30 days Regional mea
9 % Hospitalization with LOS > 30 days Regional mea
10 65 year
11 nal Mea
Table 3
Method of val
No. Interv
1 Gener
2 Paedi
3 Other
4 Pharm
5 Inapp
6 Avera
DRGs
7 Pre-su
8 Readm
9 % Hos
days
10 Poten
11 Absen
In each
number of h
able inpatie
depends on
type of inte
scenarios:
1. If the nu
variable
avoidabl
with eac
for all HA
2. If the nu
ical war
with a c
short pe
nel to ot
appears
2 This cut of
discussion on
out that perso
cut correspon
(instruments,
beds.
ger pe
ns.
the nu
or), st
mbers
lated to
ance
re bas
e resu
top H
f man
. Nota
mstan
tial saPotential retirees Over
Absenteeism rate Regio
uation applied to each indicators.
entions Method of valorization
al practitioners expenses Already expressed in
costs
atrics expenses
services expenses
aceutical expenses
ropriate hospitalization rate DRGs fares
ge length of stay for medical Use of direct costs
linked to the number of
beds that can be
re-allocated
rgical length of stay
ission within 30 days
pitalization with LOS > 30
tial retirees No. of potential
lon
cia
3. If
flo
nu
re
ten
we
Th
to all
part o
group
circu
potenretirees × average
wage
teeism rate Absenteeism
rate × total personnel
costs
case, direct avoidable costs related to the
ospital beds derives from the number of avoid-
nt bed days. The amount of resources saved
the number of beds and on the basis of the
rvention that can be planned based on three
mber of beds is lower than 12,2 only some
costs (laundry, food, pharmaceuticals, etc.) are
e. In this case, the financial value associated
h inpatient bed day is a fixed amount of D40
s.
mber of beds is some multiple of 12 (a typ-
d), reorganization of personnel is possible,
onsequent additional reduction in costs. In a
riod of time the opportunity to move person-
her services, including community-based care,
to be feasible especially for nurses, while over
f was identified by the top managers of HAs involved in the
the basis of their health information system. They pointed
nnel can be moved across the units/wards when the activity
ds to a number of beds equal to 12, while other resources
space, etc. and other fixed costs) could drop reaching 30
4. Results
Table 4
of the 2007
allocated if
or the best
sum of all t
considering
It is imp
would be m
realized ove
allocated to
performanc
the continu
hospital sta
At local
tial savings
efficient bu
example, Fi
improveme
than 9% of s
An inter
could prov
ventions. E
interventio
ment. The
ventions is
on local st
allocated b
reached Hy
it appearsian Regional median
n per DRGs Regional mean per DRGs
n Regional target
n Regional best performance
n Regional target
s Over 60 years
n Regional best performance
riods of time it may be possible to move physi-
mber of beds is some multiple of 30 (a typical
ructural interventions such as reduction in bed
are also possible with a reduction of both costs
care and fixed or overhead costs such as main-
and amortization costs. All of these calculations
ed on 2007 costs.
lts from this exploratory study were reported
A and regional managers and then reviewed as
agement training courses for this management
bly, this review did not identify any exceptional
ces or problems that would argue against the
vings.
shows that at the regional level, from 2 to 7%
D6.1 million healthcare budget could be re-
all the institutions achieve the regional average
practice. These estimates are derived from the
he inefficiencies in the 11 indicators analyzed,
the two hypotheses.
ortant to emphasize that not all of these savings
onetized immediately. Some would need to
r a longer period of time and some would be re-
other services in order to realize the improved
e. In fact, some actions such as those regarding
ity of care require cooperation of physicians,
ff, and institutions in different places.
level, large variation in the amount of poten-
in 2007, suggest that some HAs were already
t others had large room for improvement. For
g. 1 shows that under Hypothesis 2 (maximum
nt), HAs could re-allocate between 1% and more
pending.
esting finding from this simulation is that it
ide a way of setting priorities among inter-
ach HA can now distinguish areas where
ns are possible and there is room for improve-
mix of potential savings coming from inter-
different from one HA to another, depending
rengths. Fig. 2 shows the resources to be re-
y HAs for each type of intervention, if all HAs
pothesis 2 (benchmark) standard. For instance,
that TH 1 should concentrate its effort on the
S. Nuti et al. / Health Policy 95 (2010) 137–143 141
Table 4
Regional amount of resources to be re-allocated in the two hypotheses.
No. Interventions Area of interventions Hypothesis 1 Hypothesis 2
1 General practitioners expenses Primary care D 4,415,057 D 19,882,497
2 Paediatrics expenses D 2,695,250 D 9,045,799
3 Other services expenses D 9,852,081 D 35,685,081
4 Pharmaceutical expenses Pharmaceuticals D 19,506,522 D 137,932,174
5 Inappropriate hospitalization rate Hospital D 7,185,468 D 11,294,459
6 Average length of stay for Medical DRGs D 59,844,704 D 59,844,704
7 Pre-surgical length of stay D 11,658,164 D 53,369,217
8 Readmission within 30 days Continuity D 340,477 D 6,714,571
9 % Hospitalization with LOS > 30 days D 486,397 D 536,404
10 Potential retirees Human resources
management
D 6,036,021 D 60,113,119
11 Absenteeism rate D 6,442,001 D 34,162,281
Total D 128,462,142 D 428,580,305
% on the regional annual budget
6.100 million of Euro
2.11% 7.03%
Fig. 1. Percentage of resources that could be re-allocated on the 2007 budget by each HA – Hypothesis 2.
Fig. 2. Resources to be re-allocated by HA for each interventions – Hypothesis 2.
142 S. Nuti et al. / Health Policy 95 (2010) 137–143
average pre-surgical length of stay; its major inefficiency
with a potential pay-off of more than 38,000 bed days,
equal to 105 hospital beds or D21 million.
5. Limitati
There a
notably, lim
marks as gu
with overla
ventions. T
be optimist
best perfor
of the best
target. This
under-spen
pharmaceu
improving
problems a
or audit.
There is
example, th
the overall
amount of r
the sum of
overall hos
counting th
was used b
priorities ac
in the stud
is an overe
ber of indic
means that
produce ne
of resource
than the to
research co
detailed fin
to other pe
Other li
used to app
an over/und
human reso
the real eff
absent or re
estimation,
overestima
Another
heterogene
are already
could impa
large variab
to organiza
aspects.
Further
ing the lim
practical gu
ing the savi
system are
and what a
to identify
vices that represent a much higher return in benefits to
the population’s health. Experience in a number of juris-
dictions suggests that re-investment decisions should be
made in as transparent and evidence-based fashion as pos-
and w
4,35].
nclus
is ben
of com
ation p
shown
p. It w
HA ha
they n
e imp
ct on
. Mana
ing, no
rms, b
ventio
the re
l chan
me diffi
to arg
ve fina
ase. Th
lish p
ing a n
oweve
r are n
al wit
ve the
hmark
ts a m
rce re
mplica
oth po
public
than
practic
esourc
edium
velop
ices. A
ods pr
n and
easure
me in
://ww
some
ros th
g valu
Effecti
y in a
the fir
s enab
rces t
servicons and further developments
re a number of limits on the results, most
its resulting from the use of means and bench-
ides to achievable benchmarks and problems
p and attribution of savings to specific inter-
his means that the results presented here may
ic estimates of possible savings. In some cases,
mance (Hypothesis 2) is defined on the basis
performer instead of a national or regional
could lead to some problems such as the
ding (that could occur, for instance, to the
tical expenses indicator without maintaining or
quality) or data manipulation. These sorts of
re likely only amenable to more detailed study
also a problem of overlap across indicators: for
e reducing 30 day readmissions also reduces
hospitalization rate. In this case the total
esources that could be disinvested is less than
those savings attributable to readmission and
pitalization rate reductions. Despite the risk of
e same savings multiple times, the total amount
ecause of its power to support identification of
ross actions. In addition, the actions monitored
y concern only 11 indicators; so even if there
stimation, it may be offset by the small num-
ators involved in this exploratory study. This
an extension of the same methodology may
w opportunities for savings. Thus, the amount
s available for re-allocation is likely higher
tals reported under either hypothesis. Further
uld focus on the collection of more diverse and
ancial information to extend the analysis also
rformance indicators.
mitations could be addressed to the method
ly financial value: for instance there could be
er estimation linked to the methods applied to
urces management indicators: in these cases
ect depends upon the type of personnel that is
tired; for physicians there would be an under-
for administrative personnel there would be an
tion.
limit of the study could be the impact of the
ity on performance variability. Some indicators
adjusted for age and sex but other elements
ct on variability. It is authors’ opinion that the
ility showed in the results is much more due
tional factors than to other socio-geographical
research should work to explore ways of reduc-
itations on the study and on improving the
idance that can be given to managers pursu-
ngs targets. Once managers of the public health
aware of the resources they can re-allocate
ctions they can pursue, the next step will be
the targets for re-allocation, that is, the ser-
sible
[10,3
6. Co
Th
tern
alloc
were
ershi
their
that
manc
impa
itself
mark
perfo
inter
used
tiona
beco
ment
achie
incre
demo
build
H
pape
to de
achie
benc
resen
resou
the i
for b
in a
more
best
the r
the m
to de
pract
meth
vatio
to m
of so
(http
In
of Eu
mizin
Cost
cienc
only
it ha
resou
ity ofith the involvement of a wide set of stakeholders
ions
chmarking process changed the traditional pat-
plaints around the budgeting and resource
rocesses in Tuscany. The results of this study
to all CEOs in a meeting with the Regional lead-
orked as a warning to all CEOs who argued that
d reached the maximum level of efficiency and
eeded more money to achieve required perfor-
rovements. This study has also had a significant
the use of the performance evaluation system
gers are now more aware of the value of bench-
t only to comprehend how their organization
ut also to support the identification of priority
n areas to improve efficiency. Some HAs even
sults as a starting point for planning organiza-
ges in their institutions. In this context it has
cult for HAs that show large room for improve-
ue that they do not have enough resources to
ncial balance and that they needed a budgetary
is evidence has enabled the Tuscany Region to
retexts, eliminating the word “impossible” and
ew culture around the “possible.”
r, the types of disinvestments suggested in this
ot necessarily easy. They may force managers
h professional behaviour change in order to
savings. Publication and distribution of data
ing, within and outside the organizations, rep-
eans for managers [36] to make the case for
-allocation to improve performance. Likewise,
tions of this study may be extremely useful
licy makers and the top management of HAs
system that bases its action on cooperation
competition. Benchmarking helps to identify
es and subsequently allows measurement of
es that can be disinvested and re-allocated in
and long term. It will be important though
capacity for sharing and learning from best
lthough there are some differences with the
eviously described, the NHS Institute for Inno-
Improvement has adopted a similar approach
in financial terms the possible improvements
dicators using benchmarking available online
w.productivity.nhs.uk/).
cases, a single intervention could free millions
at can be re-allocated for other services maxi-
e for citizens through Cost Benefit Analysis or
veness Analysis largely used for enhancing effi-
variety of settings [37]. Even if this analysis is
st step in the process of resource re-allocation,
led HAs and Tuscany Region to identify those
hat could be moved without reducing the qual-
es.
S. Nuti et al. / Health Policy 95 (2010) 137–143 143
Acknowledgments
The authors wish to thank the researchers from Man-
agement and Health Laboratory for their help in the
development of this study and the top managers of Tuscan
Health System. The authors also want to thank the col-
leagues of the European Health Policy Group for the useful
discussion and the reviewers of the paper, especially Prof.
AdalsteinnBrownforhis excellent adviceduring the review
process of the paper.
References
[1] Mitton CR, Donaldson C. Health care priority setting: principles,
practice and challenges. Cost Effectiveness and Resource Allocation
2004;2:3.
[2] Mitton CR, Donaldson C. Resource allocation in health care: health
economics and beyond. Health Care Analysis 2003;11(3):245–57.
[3] Mitton CR, Donaldson C, Waldner H, Eagle C. The evolution of PBMA:
toward a macro-level priority setting framework for health regions.
Health Care Management Science 2003;6:263–9.
[4] Hauck K,
health ca
[5] Baltussen
need for
Resource
[6] Ham C. P
experien
[7] Viney R, H
ysis: a gu
Bulletin 1
[8] Mitton C
health au
Science &
[9] Wilson H
ments fo
[10] Elshaug A
on disinv
Journal o
[11] Pearson
National
ment effo
Research
[12] Northcot
between
nal 2005;
[13] Johnston
Journal o
[14] Di Frisch
Corriere d
[15] Nuti S. La
2008.
[16] Nuti S, Bo
in the m
agement
[17] Pink GH,
ating a ba
2001;27:
[18] Canadian Institute for Health Information, Government Hospital.
Association and the University of Toronto. Hospital Report ‘99: a bal-
anced scorecard for Ontario Acute Care Hospitals. Toronto: Ontario
Hospital Association; 1999.
[19] Sutter Women’s & Children’s Services. First pregnancy and delivery.
Clinical Initiative Resources. California: Sutter Health; 2001.
[20] Harrison ML, Graff LA, Roos NP, Brownell MD. Discharging patients
earlier from Winnipeg hospitals: does it adversely affect quality of
care? Canadian Medical Association Journal 1995;153:745–51.
[21] Brownell MD, Roos NP. Variation in length of stay as a measure of effi-
ciency in Manitoba hospitals. Canadian Medical Association Journal
1995;152(5).
[22] Nuti S. Report 2007 – Il sistema di valutazione della performance
della sanità Toscana. Pisa: Edizioni ETS; 2008.
[23] Dittrich JE, Carrel MR. Organizational equity perceptions,
employee job satisfaction, and departmental absence and
turnover rates. Organisational Behavior and Human Performance
1979;24(1):29–40.
[24] Fisher ES, Wennberg DE, Stukel TA, Gottlieb DJ, Lucas FL, Pinder EL.
The implications of regional variations in Medicare spending. Part
1: the content, quality, and accessibility of care. Annals of Internal
Medicine 2003;138:273–87.
[25] Fisher ES, Wennberg DE, Stukel TA, Gottlieb DJ, Lucas FL, Pinder EL.
The implications of regional variations in Medicare spending. Part
: health
edicine
agan P
avoida
ith uni
2.
ibbritt D
eadmiss
linical P
orghans
nd redu
esearch
anning
ospital s
orngren
mphasis
egri-Cle
va: alcu
rogramm
inquini
iolo Vit
gement
oda V. I
iseman
c in prio
edicine
evin L, G
stablish
policy
nce. Inte
007;23(
ibbard
rmance
003:22.
urray C
HO gui
conomiSmith PC, Goddard M. The economics of priority setting for
re: a literature review. H N P Discussion Paper; 2004.
R, Niessen L. Priority setting of health interventions: the
multi-criteria decision analysis. Cost Effectiveness and
Allocation 2006;4(14).
riority setting in health care: learning from international
ce. Health Policy 1997;42:49–66.
aas M, Mooney G. Program Budgeting and Marginal Anal-
ide to resource allocation. New South Wales Public Health
995;6(4).
R, Patten S, Waldner H, Donaldson C. Priority setting in
thorities: a novel approach to a historical activity. Social
Medicine 2003;57:1653–63.
, Scott S. PBMA – its role in the future purchasing arrange-
r health care services. Health Policy 1995;33:157–60.
G, Hiller JE, Moss JR. Exploring policy-makers’ perspectives
estment from ineffective healthcare practices. International
f Technology Assessment in Health Care 2008;24(1):1–9.
S, Littlejhons P. Reallocating resources: how should the
Institute for Health and Clinical Excellence guide disinvest-
rts in the National Health Service? Journal of Health Service
Policy 2007;12:160–5.
t D, Llewellyn S. Benchmarking in UK health: a gap
policy and practice? Benchmarking: An International Jour-
12(5):419–35.
DJ. Increasing value for money in health systems. European
f Health Economics 2004;5:91–4.
ia F, Foschi P. I manager: “Più di così non si può tagliare”. Il
ella Sera, Roma; 01/10/2007. p. 3.
valutazione della performance in sanità. Bologna, Il Mulino;
nini A, Murante AM, Vainieri M. Performance assessment
aternity pathway in Tuscany Region. Health Service Man-
Research 2009;22(3):115–21.
McKillop I, Schraa EG, Preyra C, Montgomery C, Baker R. Cre-
lanced scorecard for hospital system. Health Care Finance
1–20.
2
M
[26] M
in
w
4
[27] S
r
C
[28] B
a
R
[29] H
h
[30] H
e
[31] N
ti
P
[32] C
M
a
[33] C
[34] W
li
M
[35] L
E
to
e
2
[36] H
fo
2
[37] M
W
Eoutcomes and satisfaction with care. Annals of Internal
2003;138:288–98.
, Otero A, Alberquilla A, Ribera JM. Geographic variations
ble hospitalizations in the elderly, in a health system
versal coverage. BMC Health Services Research 2008;8:
W. Validation of a 28 day interval between discharge and
ion for emergency readmission rates. Journal of Quality
ractices 1995;15:211–20.
I, Heijink R, Kool T, Lagoe RJ, Westert GP. Benchmarking
cing length of stay in Dutch hospitals. BMC Health Services
2008:8.
BW. Length of stay benchmarking in the Australian private
ector. Australian Health Reviews 2007;31:150–8.
CT, Foster G, Datar SM. Cost accounting: a managerial
. Canada: Prentice Hall; 1997.
menti A. I costi rilevanti nelle decisioni di gestione opera-
ne situazioni tipiche in Amigoni F. Misurazioni d’azienda.
azione e controllo. Milano: Giuffrè Editore; 1988.
L. La gestione dei costi in ambienti produttivi “snelli” in
ali P. Strumenti per l’analisi dei costi. Percorsi di cost man-
Vol III. Torino: Giappichelli Editore; 2003.
costi di produzione. Milano: Giuffrè Editore; 1968.
V, Mooney G, Berry G, Tang KC. Involving the general pub-
rity setting: experiences from Australia. Social Science &
2003;56:1001–12.
oeree R, Sikich N, Jorgensen B, Brouwers MC, Easty T, et al.
ing a comprehensive continuum from an evidentiary base
development for health technologies: the Ontario experi-
rnational Journal of Technology Assessment in Health Care
3):299–309.
JH, Stockard J, Tusler M. Does publicizing hospital per-
stimulate quality improvement efforts? Health Affairs
JL, Evans DB, Acharya A, Baltussen RMPM. Development of
delines on generalized cost-effectiveness analysis. Health
cs 2000;9:235–51.