Orthopaedics & Traumatology: Surgery & Research (2010) 96, 113—123

ORIGINAL ARTICLE

Cost analysis in total hip arthroplasty: Experience of

a Teaching Medical Center located in Paris
T. Lernout a, C. Labalette a,d, L. Sedel c, P. Kormann b, C. Duteil a,
A. Le Divenah a, D. Bertrand a,d, S. David d,e, C. Segouin a,d,f,∗

a
Public Health and Health Economics Dept, Lariboisière—Fernand-Widal Hospital Group, Paris Hospitals Trust (AP—HP),
2, rue Ambroise-Paré, 75475 Paris cedex 10, France
b
Financial Affairs Dept, Lariboisière—Fernand-Widal Hospital Group, Paris Hospitals Trust (AP—HP),
2, rue Ambroise-Paré, 75475 Paris cedex 10, France
c
Orthopedic Surgery Dept, Lariboisière—Fernand Widal Hospital Group, Paris Hospitals Trust (AP—HP),
2, rue Ambroise-Paré, 75475 Paris cedex 10, France
d
Denis Diderot Paris-7 University, Medical School, 16, rue Henri-Huchard, 75870 Paris cedex 18, France
e
Physicians’ In-service Training Dept, HR Dept, AP—HP, 2-4, rue Saint-Martin, 75004 Paris, France
f
Wilson Center for Research in Education, Medical School, Toronto University, Toronto General Hospital,
Elizabeth Street, Toronto, Ontario, Canada

Accepted: 17 July 2009

KEYWORDS
Summary
Diagnosis-related
Background: Since the beginning of 2008, the implementation of a 100% activity-based payment
groups;
system, has made efficiency one of the prime concern for the French health-care providing
Cost analysis;
institutions. We therefore assessed the real cost of a scheduled total hip replacement (THR) in
Efficiency;
a teaching hospital and compared findings with French national data (and with the Government
Total hip replacement
Healthcare Insurance System allowance).
Hypothesis: The study should suggest possible means to optimize organization of management
and/or clinicians’ practice.
Material and methods: This is a retrospective full-cost economic study. Patients were included
only if fulfilling the following criteria: admitted in 2006; classified in Diagnosis-Related Group
(DRG) 08C23 V or 08C23W (respectively THR without and with associated comorbidity); treated
in a single department; admitted from home; and having undergone a THR (coded as NEKA020 in
the french CPT) that same year. Treatment-cost was established on the basis of data collected
from two main sources: the Information Systems Medicalization Program (ISMP) data-base, and
the finance department data, which were taken into account in line with the French National
Costs Study (NCS) structure.

∗ Corresponding author.
E-mail address: [email protected] (C. Segouin).

1877-0568/$ – see front matter © 2009 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.otsr.2009.07.008
114 T. Lernout et al.

Results: The methodology employed here follows the 2006 National Costs Scale structure. Treat-
ment costs (excluding the cost of implantable medical devices or IMDs) were estimated at
D 8,104.72 for DRG 08C23W and D 7,529.19 for DRG 08C23 V. These figures were higher than the
rates authorized in 2006 (excluding IMDs), which were D 7,677.92 for 08C23W and D 6,358.97
for 08C23 V (taking the 7% geographic coefficient into account) and than the 2005 NCS figures
(excluding IMDs) of respectively D 7,536.13 and D 6,083.59.
Discussion: Clinical units and departments need to be able to assess costs for the patholo-
gies they treat, as health-care institutions have to balance their expenditure against their
income, which largely comes from their hospital-care activity. The methodology put forward
here, of cost comparison according to the NCS structure, enables the total cost to be known.
Comparing results (expenditure line by expenditure line) against national data, selectively high-
lights the areas in which efficiency can be improved. The exactitude of the obtained results
remains, however, limited by the rules currently in use at each individual hospital’s accounting
department.
Level of evidence: Level IV, retrospective economic and decision analysis study.
© 2009 Elsevier Masson SAS. All rights reserved.

Introduction technical platforms are above the national average, there

is nothing to show whether this finding concerns the whole
As part of the financial reform package (activity-based range of pathologies managed in the department, section or
funding [ABF]), French health-care establishments (medi- hospital or just certain specific pathologies. Since clinical
cal, surgical and obstetric [MSO]) are having to implement departments rarely have any time to organize a review of
new managerial approaches. Notably, they need to know their files on such a topic, such a finding will probably not
their own production costs, as they are required to draw lead to anything actually being done.
up an annual Provisional statement of income and expendi- The present study sought to lay the foundations for
ture (PSIE). The Lariboisière—Fernand-Widal hospital group, a simple, reproductible methodology for assessing total
a Paris teaching hospital, faces the same challenges as all management costs for a given pathology on the basis
other French public health-care establishments with respect of existing medico-administrative data. The aim was to
to its MSO activity. enable hospital departments that do not have NCS-type
As far as income is concerned, each admission is classi- analytic accountancy to know the costs of managing the
fied in a Diagnosis-related group (DRG) by the Information main pathologies they deal with and make comparisons
Systems Medicalization Program (ISMP). Each DRG has a with national mean values. Implementing the approach
nationally established price attributed to it (with weight- would first of all enable the costs involved in the depart-
ing in certain geographical regions), corresponding to the ment/hospital’s main activities to be calculated. Secondly,
payment the establishment can receive for the stay [1]. the professionals concerned could then identify the cost-
Prices are set at national level, based on the data of the lines on which marginal efficiency might be achieved in
National Costs Study (NCS), conducted by some 50 volunteer terms of treatment organization. Detailed information and
institutions. comparison to national-level data form a preliminary phase
Since ABF came to be implemented across the board in prior to reflection concerning treatment organization and
2008, more than half of the income of hospitals has come practices. The final objective is to improve efficiency. Such a
directly from admissions. study might also improve cooperation between management
To balance their finances, establishments also need to and physicians [3].
forecast and, if required, control their costs. However, the Total hip replacement (THR) was selected as focus,
accountancy practiced in French public health-care estab- firstly, as being frequent, at about 12.9% of our ortho-
lishments is not such as to allow fine-grained analysis of pedic department’s activity in 2007 (DRG of THR with or
the costs entailed by treating patients and pathologies. without associated comorbidity); and secondly, in the light
There is a tool, called the Case-Mix Costs table (CMCT), of population aging [4] and of our section’s specialization
which, according to the Hospital Expertise and Audit Mis- in prosthetics in young patients, as offering considerable
sion (MEAH), ‘‘enables the MSO activity costs of a given development potential.
health-care establishment to be compared to a virtual
establishment with exactly the same activity measured in
terms of DRGs but with production costs calculated from Material and methods
the unit costs identified in the NCS’’ [2]. Applying this
at the level of the hospital, section or department allows Type of study and population
cost lines diverging from the national average to be high-
lighted. The degree of information compaction, however, is For the purposes of the study, we calculated the total
too high to allow clinicians to design any concrete improve- cost [5] of a hospital stay, from the point of view
ment plans. If, for example, costs incurred by the use of of the hospital. The study was retrospective, con-
Cost analysis in total hip arthroplasty: Experience of a Teaching Medical Center located in Paris 115

cerning patients admitted for THR in the orthopedic • type of surgery: unilateral THR without replacement of
surgery department of our hospital group during the year prosthesis, graft, osteotomy or reconstruction. The item
2006. In the 10th DRG classification, such admissions chosen for the study was thus ‘‘Hip joint replacement by
come under DRG 08C23 V or 08C23W (respectively, ‘‘Hip total prosthesis’’, coded NEKA020 on the 2nd version of
replacement without associated comorbidity’’ and ‘‘Hip the Common Classification of Medical Procedures (CCMP);
replacement with associated comorbidity’’). As this par- • type of stay: single ward; the entire stay was within the
ticular DRG classification was implemented in March 2006, same medical ward;
the study period was limited to the last 9 months of that • type of admission: patients admitted from Emergency or
year. by transfer were excluded, and only those admitted from
DRGs are meant to be homogeneous in terms of home (‘‘Home Admission Mode’’) were included.
resources, whereas there are a number of differ-
ent item codes for THR. To obtain a homogeneous Data collection
treatment cost, we therefore made the calculation
for a ‘‘standard’’ patient group, defined by three The study data-base was constructed from 3 computer-
criteria: ized data-bases of the Paris Hospitals Board (Assistance

Table 1 Work units (WU), data and calculation formulae per expenditure-line (NCS model) (excl. blood products).

Rubric WU Data Means of calculating mean stay cost

Salaries: medical staff, nurses, Hospital Invoicable day (ID) (expenditure/n ID) × MSD Finances
auxiliaries, etc day (HD) Personnel expenditure
Maintenance amortization, HD ID
medical logistics Personnel expenditure
Medical and non-medical theater Relative Personnel expenditure (expenditure/total RCI) × target RCI
staff cost index N RCIs produced by
(RCI) theater (total and
target)
Theater RCI Expenditure on
Other expenses personnel, operating
(excl. prosthesis),
medical acts, medical
logistics, gen. logistics.
N RCIs produced by
theater (total and
target)
Laboratory Bio and N Bio and Path tests (IC/total B or P) × (target Bio or Path)
Path requested (total and
target)
Imaging, functional exploration, RCI N RCIs requested (total (IC/total RCI) × target RCI
anesthesia + postanesthesia care and target)
unit
Other medicotechnical items Other medical services (IC/total RCI) × target RCI
(excl. rehab)
N RCIs requested (total
and target)
Inc. rehab AMC N AMC and AMS (IC/total AMC + AMS) × (target target)
AMS requested (total and
target)
Medical consumables (inc. IMDs), Days Expenditure on drugs, (expenditure/n ID) × MSD Finances
drugs medical consumables
ID
Target IMD expenditure
Catering, laundry, general logistics Days Expenditure per post (expenditure/n ID) × MSD Finances
ID
Overheads Days LRB overheads (expenditure LRB/n ID for all
expenditure LRB) × MSD Finances
LRB IDs
IDs
N: number; HD: hospital day; MSD: mean stay duration; ID: invoicable day; RCI: relative cost index; AMC and AMS: refer to rehab
procedures by physiotherapists; IMD: implantable medical device; IC: induced charges (induced in one dept by another); total and
target: number of all patients treated and number of target population patients treated; LRB: Lariboisière.
116 T. Lernout et al.

publique—Hôpitaux de Paris [AP—HP]): administrative data Cost of care in the hospital group and comparison
on hospital patients (GILDA© identity server), medical pro- to NCS values
cedures performed while in hospital (AR CCAM® total
procedures server), and ward reports (SIMPA© grouper). Results are shown in Tables 2 and 3. The monetary unit is
Financial data on expenditure by the hospital group and the Euro. Care costs were D 7,529.19 for DRG 08C23 V and
the AP-HP were edited, for 2006, by the Financial Affairs D 8,104.72 for DRG 08C23W, excluding the IMDs. Comparison
Department of the AP-HP. with NCS data revealed the following differences:

Calculation of hospital stay costs • the mean hospital stay was shorter than in the NCS, for
both DRG 08C23 V and DRG 08C23W (respectively, 8.8 vs
The two DRGs were distinguished for purposes of cost calcu- 12.6 and 9.3 vs 16.9 days). According to type of discharge,
lation. The itemization was that used in the 2005 National mean stay was 2 days longer for patients referred to a
Costs Study, as found in the present Results Tables. This rehabilitation center (42% of patients) than for those dis-
means that catering, laundry and maintenance costs were charged home. Even in case of rehabilitation referral,
included, unlike in many other studies of hospital stay costs mean stays were still shorter than in the NCS: respec-
[6,7,8]. tively, 10.3 and 10.0 days for DRGs 08C23 V and 08C23W;
Patients were admitted to the orthopedics department • mean cost per patient was lower (by > 5%) than in the NCS
and managed in the in-patients ward and in theater. Costs in three areas: personnel, imaging and overheads;
for both of these sectors were analyzed, and calculated by • mean catering and laundry costs per patient were compa-
the most appropriate Work Unit (WU). Table 1 presents the rable to NCS values for DRG 08C23W, as were mean
WUs, the data and the calculation for each cost item. Cost catering and general logistics costs for DRG 08C23 V, while
per day was calculated in terms of the number of invoica- overheads costs were lower for both;
ble administrative hospitalization days (used for financial • in contrast, mean medical logistics costs per patient (line
data), so that the data would be homogeneous. Thus, each 5) were significantly higher than in the NCS; smaller dif-
‘‘ISMP’’ stay duration was increased by 1 day to obtain the ferences were also found for theater costs (line C) and
administrative stay duration. ‘‘Total medico-technical’’ costs (line D);
Certain NCS cost-lines were not used in the study: obstet- • finally, there was a difference of 15% or 25% (respectively,
ric theater, dialysis, Emergency and mobile emergency and for DRGs 08C23W and 08C23 V) for consumables, medica-
mobile intensive care unit (SMUR), intensive care, and exter- tion and blood costs (line E).
nal procedure costs. Our target cases were non-emergency
single ward stays, and there was thus no involvement of
departments other than orthopedics. Thus, care costs for our target DRGs (08C23W and
Results for the 2 DRGs were compared to the costs given 08C23 V) were higher than those found in the NCS and than
by the NCS and the rates applied by the French National the CPAM rate after application of the geographic coeffi-
Health Insurance scheme (CPAM). cient.

Results
Discussion
Target population
Methodology
In all, 57 admissions under DRG 08C23 V and 72 under
08C23W were included (Fig. 1). The methodology used to calculate care costs was based
on the cost-structure of the 2005 French National Costs
Scale. A Medline review was carried out in June 2008,
using the MeSH keywords ‘‘Arthroplasty, Replacement’’,
‘‘Cost and Cost Analysis’’, and ‘‘Arthroplasty, Replacement,
Hip/economics’’, to compare our methodology to others. It
found no published studies with detailed calculation of total
hospital THR management costs. Chamberlin et al. [6] deter-
mined the direct hospital cost of managing pertrochanteric
fracture in the elderly, and Stargart [9] assessed differ-
ential primary THR costs across 9 EU member countries,
but with insufficient methodological detail for comparison
with the present results. Finally, the Rhône-Alpes Regional
Union of Private Physicians (URMLRA) in France compared
costs between three surgical procedures [10]. All in all,
whether before or after the introduction of ABF, few such
studies have been published. [4,5], although some of the
Figure 1 Target population with (08C23V) and without cost-calculation methods presented in the above-mentioned
(08C23W) comorbidity for scheduled THR admission in 2006. MEAH report do show certain similarities to ours.
 Cost analysis in total hip arthroplasty: Experience of a Teaching Medical Center located in Paris
Table 2 Full unit cost (Euros) of DRG 08C23W in 2006.

Lariboisière (NEKA 020) NCS 2005a Diff. HG/NCS

HG expenditure
NCS (expenditure)/NCS
(expenditure)

Number of admissions or 72 Number of admissions or 1 365

sessions sessions
Lariboisière ISMP MSD 9.3 NCS MSD (days) 16.9
(days)
MSD Finance (days) 10.3 NCS IC stay (days) 0.1
Mean age (years) 63.7
1 Medical staff salaries 138.41 Medical staff salaries 376.14 −63%
2 Nurse + auxiliaries 1565.68 Nurse salaries (inc. 1699.60 −8%
salaries auxiliaries)
3 Other salaries 332.71 Other salaries 431.47 −23%
A=1+2+3 Total personnel 2036.80 Total personnel 2507.21 −19%
4 Maintenance 18.18 Maintenance 20.74 −12%
amortization amortization
5 Medical logistics 576.84 Medical logistics 118.23 388%
B=4+5 Maintenance 595.02 Maintenance 138.97 328%
Amortization + medical Amortization + medical
Logistics Logistics
6 Medical theater staff 234.65 Theater staff 682.04 −9%
7 Non-medical theater 384.10
staff
8 Other theatre 1228.64 Other theatre 404.71 204%
expenditure expenditure
C=6+7+8 Total theater 1847.38 Total theater 1086.75 70%
9 Lab 517.82 Lab 253.18 105%
10 Obstetric theater 0.00 Obstetric theater 0.91 NA
11 Imaging 44.99 Imaging 117.44 −62%
12 Functional exploration 3.32 Functional exploration 20.79 −84%
13 Dialysis 0.00 Dialysis 21.07 NA
14 Anesthesia + postanesthesia 853.02 Anesthesia + postanesthesia 682.30 25%
care unit care unit
15 Emergency 0.00 Emergency 41.11 NA
16 SMUR 0.00 SMUR 5.41 NA
17 X-ray 0.00 X-ray 2.45 NA
18 IC IC 139.95 NA
19 Other medico-technical 472.17 Other medico-technical 58.33 709%
inc. rehab 472.17

117
 118
Table 2 (Continued )

Lariboisière (NEKA 020) NCS 2005a Diff. HG/NCS

HG expenditure
NCS (expenditure)/NCS
(expenditure)

D = 9 to 19 Total medico-technical 1891.32 Total medico-technical 1342.94 41%

procedures procedures
Medical consumables 1769.99 Medical consumables 1117.74
(inc. IMDs) (inc. IMDs)
Blood products Blood products 250.75
Drugs Drugs 173.98
E Total consumables, 1769.99 Total consumables, 1542.48 15%
drugs, blood drugs, blood
20 External procedures External procedures 60.46
I = A to E + 20 Total ‘‘medical’’ cost 8140.51 Total ‘‘medical’’ cost 6678.79 22%
II Catering 226.92 Catering 209.71 8%
III Laundry 78.54 Laundry 84.45 −7%
IV General logistics 1182.71 General logistics 1374.65 −14%
V Overheads 246.02 Overheads 306.26 −20%
Total cost inc. 9874.71 Total cost inc. 8653.87 14%
Overheads Overheads
Total cost inc. 8104.72 Total cost inc. 7536.13 7.5%
Overheads, excl. IMD Overheads, excl. IMD
MSD: mean stay duration; IMD: implantable medical device; NCS: national costs scale; DRG: diagnosis-related group; HG: hospital group; SMUR: Mobile intensive care team (service mobile
d’urgence et de réanimation); IC: induced charges (induced in one dept by another).
a The 2005 National Costs Scale was used, as it set rates for 2006.

T. Lernout et al.
 Cost analysis in total hip arthroplasty: Experience of a Teaching Medical Center located in Paris
Table 3 Full unit cost (Euros) of 08C23 V in 2006.

Lariboisière (NEKA 020) NCS 2005a Rapport GH/ENC

Dépenses (GH)
Dépenses (ENC)/(Dépenses
ENC)

Number of admissions or 57 Number of admissions or 7069

sessions sessions
Lariboisière ISMP MSD 8.8 NCS MSD (days) 12.6
(days)
MSD Finance (days) 9.8 NCS IC stay (days) 1.1
Mean age (years) 62.7
1 Medical staff salaries 131.70 Medical staff salaries 263.50 −50%
2 Nurse + auxiliaries 1489.68 Nurse salaries (inc. 1426.38 4%
salaries auxiliaries)
3 Other salaries 316.56 Other salaries 362.58 −13%
A=1+2+3 Total personnel 1937.93 Total personnel 2052.46 -6%
4 Maintenance 17.29 Maintenance 14.38 20%
amortization amortization
5 Medical logistics 548.84 Medical logistics 97.34 464%
B=4+5 Maintenance 566.13 Maintenance 111.72 407%
amortization + medical amortization + medical
logistics logistics
6 Medical theater staff 230.27 Theater staff 628.80 −3%
7 Non-medical theater 376.93
staff
8 Other theatre 1205.72 Other theatre 356.70 238%
expenditure expenditure
C=6+7+8 Total theater 1812.92 Total theater 985.50 84%
9 Lab 323.74 Lab 149.38 117%
10 Obstetric theater 0.00 Obstetric theater 0.00 NA
11 Imaging 41.43 Imaging 79.30 −48%
12 Functional exploration 0.89 Functional exploration 7.80 −89%
13 Dialysis 0.00 Dialysis 0.17 NA
14 Anesthesia + postanesthesia 745.21 Anesthesia + postanesthesia 579.01 29%
care unit care unit
15 Emergency 0.00 Emergency 25.90 NA
16 SMUR 0.00 SMUR 2.03 NA
17 X-ray 0.00 X-ray 1.37 NA
18 IC 0.00 IC 7.18 NA
19 Other medico-technical 450.92 Other medico-technical 40.44 1015%
Inc. rehab 450.92

119
 120
Table 3 (Continued )

Lariboisière (NEKA 020) NCS 2005a Rapport GH/ENC

Dépenses (GH)
Dépenses (ENC)/(Dépenses
ENC)

D = 9 to 19 Total medico-technical 1 562.20 Total medico-technical 892.58 75%

procedures procedures
Medical consumables 1865.73 Medical consumables 1222.14
(inc. IMDs) (inc. IMDs)
Blood products Blood products 185.25
Drugs Drugs 85.01
E Total consumables, 1865.73 Total consumables, 1492.40 25%
drugs, blood drugs, blood
20 External procedures External procedures 29.41
I = A to E +20 Total ‘‘medical’’ cost 7744.91 Total ‘‘medical’’ cost 5564.07 39%
II Catering 215.90 Catering 182.58 18%
III Laundry 74.73 Laundry 69.83 7%
IV General logistics 1 125.30 General logistics 1 187.55 −5%
V Overheads 234.08 Overheads 256.70 −9%
Total cost inc. 9394.92 Total cost inc. 7260.73 29%
Overheads Overheads
Total cost inc. 7529.19 Total cost inc. 6083.59 24%
Overheads, excl. IMD Overheads, excl. IMD
MSD: mean stay duration; IMD: implantable medical device; NCS: national costs scale; DRG: diagnosis-related group; HG: hospital group; SMUR: Mobile emergency team (service mobile
d’urgence et de réanimation); IC: induced charges (induced in one dept by another).
a The 2005 National Costs Scale was used, as it set rates for 2006.

T. Lernout et al.
Cost analysis in total hip arthroplasty: Experience of a Teaching Medical Center located in Paris 121

Results the human and material resources directly deployed in each

CCMP procedure [11]. Our choice was due to the presence of
The cost differentials found here may have several causes. such lines in the analytic accounts system we had available.
Firstly, we included only a certain number of the patients Moreover, our analytic accounts system fails fully to distin-
falling under our target DRGs. There are also other possible guish IMDs paid for as extras (nails, cement, etc.), which may
causes, which we shall examine by cost-sector. have increased the mean RCI cost. Furthermore, the signifi-
cant difference in medico-technical procedure costs (line D)
may be due to our having included rehabilitation procedures
Medical staff expenditure in the ward
under line 19 (‘‘Other medico-technical’’) rather than line
The difference observed here is due to the combination
A (personnel expenditure).
of two factors. The first concerns how time is shared out
Not all cost analyses take account of general expendi-
between theater, consultations and the in-patients depart-
ture [6,7,8], although a useful database for this exists (the
ment. In our institution, each head of department draws
Angers data-base). Some studies, moreover, do not include
up this distribution yearly for accountancy purposes, and
a calculation of investment and amortization costs [8]. We
it is not then objectified by any precise measurement of
chose to take these on board, as the PSIE was drawn up
work-time spent on practitioners’ various activities. More-
with these two elements taken into account. However, the
over, in this teaching hospital, the department is staffed
specificity of the AP—HP, of which our hospital group is a
by university-employed physicians, who cost the institution
part, must be borne in mind: HQ costs are spread over the
less than hospital-employed practitioners would: the hos-
36 institutions making up the AP—HP. Expenses reduced to
pital itself pays a university professor less than 60% of the
the ‘‘day of hospital stay’’ WU are strongly affected by the
salary it pays to a full-time hospital practitioner.
duration of the stay. The present mean stay durations were
shorter than in the NCS and may mask mean daily stay costs
Medical expenditure (laboratory, imaging and that were higher. Also, the difference in mean stay duration
consumables) according to type of discharge (home or rehab center) is
The total laboratory cost differential may have been due to worth examining in a dedicated study: the present research
the number of laboratory acts being higher than the national did not distinguish the two, and the issue is being addressed
average, or to higher production costs. It therefore has to as part of the clinical trajectory analysis being undertaken
be determined whether this was a volume effect or a price in the orthopedics department at the time of writing.
effect. If it was a volume effect or prescriber effect (e.g., Finally, the total costs for the two DRGs were almost
trainee physicians), this needs discussing with the clinicians identical, whereas 08C23W might have been expected to be
as it is a question of practices. Imaging costs, in contrast, more costly. Several hypotheses would be worth exploring,
were half those of the NCS, for which there are two possi- notably the relatively short duration of stays for 08C23W in
ble explanations. The available NCS data fail to distinguish our institution, despite perfectly standard management (the
between scheduled THR costs and those for THR secondary exactitude of the DRG coding having been double-checked).
to fracture, which require more X-ray items. Moreover, X- Further information will be needed in order to guide action.
ray production costs may be lower in our institution than
in the NCS institutions, or again our department’s protocol
may entail fewer X-rays. These points are currently under Interest of the study
study.
Finally, the cost differential in consumables, drugs and Working from existing data, we sought to determine the cost
blood products (line E) may be due to the use of specific of a procedure that is homogeneous and frequent in our
and more costly IMDs for younger patients (63.7 and 62.7 institution. Comparison between our calculated costs and
years, respectively for DRGs 08C23W and 08C23 V). It is to those of the NCS and the CPAM rate [11] revealed certain
be noted that lack of available accountancy data prevented differences. These findings, in turn, lead us to explore the
our costing the blood products actually used for the study reasons for such differences, in close collaboration with the
patients; a mean cost for orthopedic department patients departments concerned, whether administrative, medico-
as a whole was used instead. technical or clinical.
The approach by cost-line helps target corrective action
Other expenditure as part of the orthopedic department’s quest for efficiency.
The significant difference in medical logistics costs (line 5) Improved patient management should help optimize costs;
compared to the NCS values may be due to how they are Greater awareness of the interest of early standing and
apportioned by AP—HP HQ, so that catering, laundry, gen- the precautions to be taken by the care-team, or again
eral logistics and overheads vary from one establishment to preplanned and more frequent intervention by physiother-
another. On-going amortization of a new sterilization system apists, should combine to reduce mean hospital stay. This
may also enter into this differential. would reduce those costs that are directly bound to stay
The theater costs differential (line C) may correspond to duration. Moreover, this activity-based rather than DRG-
a significant potential for productivity gains by optimizing based approach is closer to that of the clinician, enabling
theater occupation (as confirmed by a subsequent audit). physicians to be more actively involved in the search for
Moreover, our accounting procedure included theatre logis- solutions. This kind of approach may help guide the choice
tics under the WU ‘‘relative cost index’’ (RCI), although this of issues in the assessment of professional practices: e.g.,
was inappropriate to the definition of RCI as the accountancy the relevance of prescribing complementary imaging and
WU for medico-technical procedures. The RCI represents biological examinations.
122 T. Lernout et al.

Drawing up ‘‘clinical trajectories’’ [12] could help opti- The first is to relate data on activity and on expenditure
mize management. Depending on the presenting pathology, when the ISMP and finances are not cut up in the same
a patient’s care-trajectory within the establishment is way. Also, for certain items, our hospital group is dependent
determined as of admission. This presenting pathology, how- on how the AP-HP spreads its costs. Finally, some technical
ever, is not the sole relevant factor, as possible comorbidity difficulties were encountered in using the various software
and medico-psycho-social factors influence the indications programs, which leads us to call for using a single integrated
for procedures and expert involvement along the trajectory, package for all aspects of patient management.
all of which impact stay duration. Such factors need to be
taken into account at admission.
The study had the virtue of promoting a genuine part-
Conclusion
nership between the Public Health and Health Economics
Department, the clinical department and the Finance Activity-based funding means that public health-care
Department, and this at a time when efficiency is being establishments need to evolve from expenditure-control
explicitly required of health-care establishments by the management to cost-control management. Given the
authorities. This work needs refining, with more precise cal- national rate applied to each stay, establishments need to
culation of personnel and consumables costs, which it is know their production cost, at least for their main activi-
possible to detail per patient. ties (in terms of volume and/or resources deployed), and
especially those for which efficiency might be improved. The
present study was conducted with this in mind. It is then up
Study limitations to those involved to work together with efficiency as their
aim. We were able to highlight differentials with respect to
In the terms of the review of the medico-economic literature national averages, excluding IMDs (7.5% in 08C23W and 24%
conducted by Bozic et al. [13], the present study corre- in 08C23 V). The 7% differential with respect to the weighted
sponds to the 58% of cost-identification studies and the 80% CPAM rate (respectively −2% and −5.5%) was better, but
carried out from the point of view of the establishment. The too slight to be interpretable, given the above-mentioned
results are to be taken with caution, as tendencies. Unlike study limitations. We are all too well aware of the fact
certain cost analyses, we did not calculate either the hourly that the lack of precise analytic accountancy data and the
costs of medical and paramedical staff [11] and their distri- approximations which that imposed on us precludes any
bution between theater and ward or exact costs for medical clear assertion as to whether THR costs in our hospital group
consumables and imaging [6,7,14] or laboratory examina- are higher or lower than the national average, especially as
tions for target patients. It is worth noting that medical and regards the heavier cases classified as 08C23W. The present
paramedical activities are concentrated in the days directly study did, however, manage to raise a certain number of
following surgery. In other words, THR costs are not evenly questions, highlighting the complexity of assessing treat-
spread over the stay. The Work Unit chosen for personnel ment costs. Our hospital group has decided to take part in
costs in the present study is thus open to reflection. More- the NCS, which should help provide more useful information.
over, the distribution of medical personnel costs between The study also helped identify points to which special atten-
theater and ward depends on the distribution of medical tion should be paid in order to set up a process to improve
personnel’s work-time between the various sectors (consul- THR patient management. It has been decided to draw up a
tation, theater and ward), which is self-reported. clinical THR trajectory. This will be followed by a new costs
The study was also based on data for activities as coded analysis, to assess what progress has been made.
by the physicians themselves. The quality of this information
depends upon exactitude and thoroughness of the coding. In
2006, the thoroughness of ISMP coding reached 100% in the Conflicts of interest
orthopedics department.
Moreover, not all cases classified under the two DRGs None, for all authors
were analyzed. Only ‘‘single department’’ stays were
included, so as to have precise costs for the orthope- Acknowledgments
dics department (theater and ward) alone, cost distribution
being more complicated when patients change departments
Pascal Pierson, Jean-François Tacnet, Anne-Christine Gré-
during their stay. Likewise, non-scheduled admissions were
goire, Valérie Barbier, Fabien Martinez, Didier Brosillon and
excluded, and their costs may tend to be higher. And finally,
Alexandrine Ferrand
the DRGs 08C23 V and 08C23W were set up in March 2006,
so that our studied concerned only the last 9 months of
that year. All of these choices, however, had the advantage References
of analyzing costs on well-defined and homogeneous care
procedures. [1] Segouin C, Bréchat PH, Rymer R. Tarification à l’activité: des
défis pour 2006. Presse Med 2006;35:565—7.
[2] Mission d’expertise et d’audit hospitalier (MEAH). ‘‘Nouvelle
Difficulties encountered
gouvernance et comptabilité analytique par pôle’’.
http://www.meah.sante.gouv.fr 2006.
Many difficulties can be expected during costs analysis, [3] Pirson M, Leclercq P. Comparaison du coût de la prise en charge
especially if the institution is not involved in the National de l’infarctus du myocarde entre trois établissements hospital-
Costs Study or does not practice fine-grained accountancy. iers. J Econom Med 2005;23:439—55.
Cost analysis in total hip arthroplasty: Experience of a Teaching Medical Center located in Paris 123

[4] Birell F, Johnell O, Silman A. Projecting the need for hip [14] Gerbaud L, Chiambaretta F, Desrumeaux H, et al. Cost effi-
replacement over the next three decades: influence of chang- ciency study of lacrimal canal obstruction treatment. J Radiol
ing demography and threshold for surgery. Ann Rheum Dis 2003;84:41—6.
1999;58:569—72.
[5] Launois R, Vergnenègre A, Garrigues B. Notions et mesures
des coûts en fonction de la perspective choisie. Bill Cancer
2003;90:946—54. Glossary
[6] Chamberlin B, Laude F, Rolland E, Langer H, Saillant G. Evalua-
tion of the direct cost of trochanteric fractures in the elderly. Terms for cost analysis in common use in France ABR: activity-
Rev Chir Orthop 1997;83:629—35. based rates (tarification à l’activité: T2A)
[7] Chamberlin B, Benazet JP, Humbert B, Saillant G. Evalua- AP-HP: Assistance publique—Hôpitaux de Paris: Paris Teaching Hos-
tion du coût direct des fractures du poignet. Rev Chir Orthop pitals Administrative Trust
1999;85:828—33. CCMP: Common Classification of Medical Procedures (Classification
[8] Guillemin F, Cao MM, Girard F, et al. Cost of medical commune des actes médicaux: CCAM) equivalent to Common
imaging practices in acute abdominal syndromes. Press Med Procedural Terminology (CPT)
2000;29:829—34. CMCT: case-mix costs table
[9] Stargardt T. Health service costs in Europe: cost and reimburse- CPAM: Caisse primaire d’assurance maladie: French National Health
ment of primary hip replacement in nine countries. Health Econ Insurance Third-Party Payer
2008;17(1 Suppl.):S9—20. DRG: diagnosis-related group (groupe homogène de malades: GHM)
[10] Caton J, Colin C, Fanani F. Coûts comparés de trois interven- IMD: implantable medical device
tions chirurgicales. 2004. Étude disponible sur le site de l’Union ISMP: Information Systems Medicalization Program (Programme de
Régionale des Médecins Libéraux Rhône-Alpes: www.urmlra. médicalisation des systèmes d’information: PMSI)
org/upload/urmlra/urm etude/pj/3INTERVENTIONS.pdf, MSO: medical—surgical—obstetric
2009. NCS: National Costs Study/National Costs Scale (Étude/Échelle
[11] El Khatib K, Danino A, Malka G. Étude de la corrélation entre nationale de coût: ENC)
le coût réel d’une intervention et ses modes de facturation en PSIE: provisional statement of income and expenditure (état prévi-
France: exemple de l’ablation de matériel d’ostéosynthèse de sionnel de recettes et de dépenses: EPRD)
mandibule. Rev Stomatol Chir Maxillofac 2004;105:143—8. RCI: relative cost index
[12] Kim S, Losina E, Solomon DH, et al. Effectiveness of clinical SMUR: Service mobile d’urgence et de réanimation: Mobile Intensive
pathways for total knee and total hip arthroplasty. Literature Care Unit
review. J Arthroplasty 2003;18:69—74. THR: total hip replacement
[13] Bozic KJ, Saleh KJ, Rosenberg AG, Rubash HE. Economic eval- WU: work unit (unité d’œuvre)
uation in total hip arthroplasty: analysis and review of the
literature. J Arthroplasty 2004;19:180—9.