Introduction
The German healthcare system has been undergoing deep structural and cultural changes
in recent years. The German legislative body has imposed a number of new general
conditions which should make medical treatments and the costs and revenues of
hospitals more transparent as well as encouraging hospitals to be more
cost-effective and profitable. Given the increased competitive pressures, the
assumption is that around 30 % of hospitals in Germany will have to close in the
next 10 years [1 ], [2 ], [3 ].
For this reason, health economic evaluations have increasingly honed in on the
finances of perinatal centres. If a centre intends to maintain its maternity
department, the first consideration must be to determine which best level of care
would best suit the hospital (levels I–IV as defined by the criteria of the German
Joint Federal Committee) [4 ]. Defining the right treatment
priorities and making decisions about the range of profitable and unprofitable
treatments offered by the centre will become increasingly important in the years to
come [5 ].
To remain competitive and improve a centreʼs profitability it will be necessary to
optimise the time patients spend in hospital. Reducing the number of acute beds is
one way of dealing with the increased cost pressures as the provision of acute beds
involves high fixed costs [6 ], [7 ]. It is essential that hospital managements are aware of the most
“cost-effective” mix of patients and procedures for their facility together with the
revenues and contribution margins [8 ], [9 ]. Numerous hospitals have become specialist centres
offering more complex procedures and “higher value” DRGs (Diagnosis Related Groups)
in the hope of achieving higher revenues [8 ]. Whether this
strategy can be recommended for perinatal centres from a health economic standpoint
requires closer scientific analysis. Level I perinatal centres are centres offering
maximum medical care and, in principle, they are not permitted to reject any cases.
Nevertheless, using certain marketing and cooperation strategies it is possible to
regulate the patient mix and influence the case numbers receiving various treatments
while offering the same range of services or, in the event of a particularly
loss-making patient mix, to negotiate specific additional charges with funding
agencies.
In addition to the actual numbers of cases, the type of DRG plays a very important
role for perinatal centres, as cases can be very heterogeneous, ranging from women
with preterm premature rupture of membranes in the 28th week of gestation who
require in-hospital bed rest to spontaneous onset of labour at term followed by
delivery without complications with the mother only remaining in hospital for one
day [10 ].
Perinatal centres are also obliged to comply with the requirements prescribed by the
Federal Joint Committee and these requirements involve additional resources and
costs. Physicians and nursing staff must be present at all times in delivery units
and emergency units; a neonatologist must always be on standby, an obstetrician
specialised in special obstetrics and perinatal medicine must always be on call.
There are also a number of guidelines on infrastructure; for example, the neonatal
intensive care unit must be in the immediate vicinity of the delivery unit and
operating room and the neonatal intensive care unit must have at least six intensive
care beds [4 ], [11 ].
These detailed and cost-intensive requirements and the current discussion about the
minimum size of facilities mean that perinatal medical care currently faces
important challenges which will require carefully considered planning and management
decisions. A detailed knowledge of the potential impact of the different treatments
offered is essential for decision-making. The following questions, which were
investigated in the present study using theoretical models, are of particular
interest when a hospital management is considering how to secure the long-term
survival of their hospital:
What is the impact of different DRGs or services and treatments on operating
profits?
Does it make more economic sense to treat more high-risk patients or is it
better to treat more patients with a lower risk profile?
What economic impact do different obstetrical procedures have; in particular,
how does spontaneous birth compare to caesarean section?
What is the effect of fully utilising all bed capacities and operating room
minutes and what will occur if some of the capacities remain unused?
Material and Methods
The data of the 2009 multistage contribution margin accounting of the University
Perinatal Centre for Franconia (UPF) were used as the basis for all theoretical
costing models [12 ].
When calculating capacity constraints, the P-DRGs (neonates) were not included in
the
capacity of 36 beds available in the obstetrical department of the University
Hospital Erlangen. Only the DRGs O01A-O40Z were included in the OP-DRGs (operative
DRGs) ([Table 1 ]).
Table 1 DRGs arranged according to profitable and
loss-generating treatments (middle grey ▒: DRGs which are profitable for
the perinatal centre, dark grey ▓: DRGs which are less profitable for
the perinatal centre, light grey ░: midway position between profitable
and less profitable DRGs = DRGs with no effect on
revenues).
DRG
Brief description of the DRG
Average CM I per DRG case per inpatient day in €
DRG = Diagnosis Related Groups; CM = contribution margin;
O01A-P67D = DRGs relevant for perinatal centres; OR
procedure = operating room procedure, significant intervention
as defined by the DRG system; CC = complications and/or
co-morbidities
O01A
Caesarean section with several complicating diagnoses, up until
the end of the 25th week of gestation or with intrauterine
therapy or complicating constellation
193.00▓▓
O01B
Caesarean section with several complicating diagnoses, 26th to
33rd completed week of gestation, without intrauterine therapy,
no complicating constellation
269.49▓▓
O01C
Caesarean section with several complicating diagnoses, > 33rd
week of gestation, without intrauterine therapy, no complicating
constellation or with complicating diagnosis
333.76▓▓
O01D
Secondary caesarean section with several complicating diagnoses,
> 33rd week of gestation, without intrauterine therapy, no
complicating constellation or with complicating diagnosis
371.89░░
O01E
Secondary caesarean section with complicating diagnosis,
> 33rd week of gestation, no complex diagnosis
376.38░░
O01F
Caesarean section without complicating diagnosis, > 33rd week
of gestation, 33rd week of gestation, no complex diagnosis
448.28▒▒
O02A
Vaginal delivery with complicated OR procedure, up to 33rd week
of gestation or with intrauterine therapy or complicating
constellation
535.39▒▒
O02B
Vaginal delivery with complicated OR procedure, > 33 SSW,
without intrauterine therapy, no complicating constellation
384.85▓▓
O03Z
Ectopic pregnancy
430.05▒▒
O04Z
Admission to hospital after delivery or miscarriage requiring OR
procedure
191.89▓▓
O05A
Cerclage and cervical cerclage
611.44▒▒
O05B
Certain OR procedures during pregnancy without cerclage, without
cervical cerclage
518.14▒▒
O06B
Intrauterine therapy of the foetus without laser therapy
263.51▓▓
O40Z
Miscarriage with dilation and curettage, aspiration curettage or
hysterotomy
387.62░░
O60A
Vaginal delivery with several complicating diagnoses, at least
one of them serious, up to 33rd week of gestation or with
complicating constellation
454.18▒▒
O60B
Vaginal delivery with several complicating diagnoses, at least
one of them serious, > 33rd week of gestation, no
complicating constellation
391.73░░
O60C
Vaginal delivery with serious or moderately serious complicating
diagnosis
376.64░░
O60D
Vaginal delivery without complicating diagnosis
416.56▒▒
O61Z
Admission to hospital after delivery or miscarriage without OR
procedure
250.26▓▓
O62Z
Imminent miscarriage
365.45░░
O63Z
Miscarriage without dilation and curettage, aspiration curettage
or hysterotomy
236.92▓▓
O64A
Ineffective contractions, more than one day of
hospitalisation
208.18▓▓
O64B
Ineffective contractions, one day of hospitalisation
282.43▓▓
O65A
Other pre-birth hospital admissions with extremely serious or
serious CC
98.76▓▓
O65B
Other pre-birth hospital admissions without extremely serious or
serious CC
176.17▓▓
P60A
Neonate, died < 5 days after admission without significant OR
procedure
2 041.92▒▒
P66C
Neonate, weight at admission 2 000–2 499 g without significant OR
procedure, no ventilation > 95 h, with other problems
1 276.11▒▒
P66D
Neonate, weight at admission 2 000–2 499 g without significant OR
procedure, no ventilation > 95 h, without problems or
multiple births, weight at admission > 2 499 g
284.11▓▓
P67B
Neonate, weight at admission > 2 499 g without significant OR
procedure, no ventilation > 95 h, with serious problem
902.76▒▒
P67C
Neonate, weight at admission > 2 499 g without significant OR
procedure, no ventilation > 95 h, with other problem
449.17▒▒
P67D
Newborn singleton, weight at admission > 2 499 g without OR
procedure, no ventilation > 95 h, no serious problem, no
other problem
203.80▓▓
Unless otherwise stated in the brief description of the individual theoretical model,
the real number of days in hospital was established for each respective DRG and the
maximum number of cases was calculated for each DRG based on the total number of
days stayed in hospital in 2009.
The total number of days stayed in hospital and the total OR (operating room) minutes
were defined in the model as the capacity constraints. The percentage each
individual DRG contributed to the total number of days in hospital and the total
number of OR minutes in 2009 was calculated. In the next step, the maximum number
of
cases which could be treated if only this DRG was treated was calculated for every
DRG based on the average number of days spent in hospital for every individual DRG,
taking the calculated capacity constraint of 13 140 hospital days into account. If
it was assumed that surgical procedures were performed in the operating room of the
UPF delivery unit for 10 hours every weekday of the year and 4 hours per day on
weekends and public holidays, the potential total OR capacity was calculated as
176 880 OR minutes. If, instead, it was assumed that surgical procedures were
carried out every day of the year, 24 hours a day, the OR capacity was 525 600 OR
minutes.
Subsequently the utilisation percentage for the capacity constraint was established
for every DRG in the different theoretical models; the capacity constraint was
assumed to be 100 %, making it possible to compare the models to one another.
Once the case numbers for the theoretical models had been established, data from the
calculations for the real model of the UPF were used again. The total number of days
in hospital was calculated by multiplying the ascertained number of cases with the
average number of days in hospital per DRG case. The total OR time for every DRG was
calculated analogously. The case numbers for every DRG were multiplied with the
average contribution margin I for every DRG case; the results were then added up and
the total was the contribution margin I.
In the next step, the UPFʼs fixed costs in 2009 were included. The calculated
contribution margin II for in-hospital (DRG) services was added to the contribution
margin II for outpatient services (prenatal outpatient clinics and pregnancy clinics
of the UPF) for 2009. These figures were not modified in the theoretical models. The
general and administrative lump sum costs amounting to 19.3 % of all UPF revenue
were deducted from the sum of the two contribution margins II.
This study did not examine the cost-revenue situation for outpatients and outpatient
services as the theoretical models only focussed on the impact of various
combinations of in-hospital treatments. The economic benefits and disadvantages of
outpatient management structures have been examined elsewhere in more detail in a
real model [12 ].
In principle, the ratio between the number of births and the P-DRG payments should
be
the same in the theoretical models as in the real model. In 2009, there were 2143
births in the perinatal centre and 1615 P-DRGs; this corresponds to a ratio of
births to P-DRGs of 1.33. The number of P-DRGs in the theoretical models was
therefore calculated by dividing the number of births in the respective model by the
ratio of births to P-DRGs calculated for the year 2009.
To adjust the operating income to the potential over- or under-utilisation of beds
it
was necessary to calculate how many beds would be required for different mixes of
cases. Personnel costs are step costs which do not increase or decrease in
proportion to bed occupancy rates but increase or decrease stepwise. Nevertheless,
for the purposes of simplification, in the theoretical models changes in fixed costs
were calculated as proportional changes based on the number of additional beds
required.
Results
The average contribution margin I per inpatient day for each DRG case can be
calculated using the average contribution margin I together with the average time
spent in hospital for each DRG. Using the respective real costs, DRGs can be divided
into profitable and loss-making DRGs. This information obtained from theoretical
models is highly relevant when considering which treatments to offer.
The most profitable DRGs were spontaneous births occurring at term and without
complications (O60D), vaginal births with complicating OR procedures (an operating
room procedure was a significant intervention as defined by the DRG system) or
several complicating diagnoses up until the 33rd week of gestation (O02A, O60A) as
well as caesarean sections (c-sections) without complicating diagnoses performed
after the 33rd week of gestation (O01F). In the UPF, the following DRGs were
classified as less profitable when the real costs were included: caesarean section
with several complicating diagnoses (O01A–O01C), vaginal delivery with complicated
OR procedure after the 33rd week of gestation (O02B) and prenatal maternal
admissions to hospital (O64A, O64B, O65A, O65B). [Table
1 ] shows the respective contribution margin I for every DRG case per
inpatient day. The brief description provides a summary of the key points for each
DRG. DRGs classed as profitable are highlighted in middle grey, unprofitable DRGs
highlighted in dark grey. All non highlighted DRGs were neither profitable nor
definitely loss-making.
A total of 19 theoretical models were compiled ([Table
2 ]). Fifteen of these models could be compared directly with each other as
they all featured bed occupancy rates of 100 %. All models were below the potential
and realistic surgical capacity limit of 176 880 surgical minutes, with the
exception of model O, which was slightly higher with 194 296 surgical minutes. As
the capacity constraints were comparable it was assumed that the fixed costs were
also virtually identical.
Table 2 Overview of theoretical models and their results
(light grey ░: real reference models [real model 2009 as well as
possible real operational results at full capacity utilisation]; dark
grey ▓: less profitable and loss-making models compared to real cost
models; middle grey ▒: more profitable models compared to real cost
models).
Model
Brief description of model
Bed occupancy (%)
Surgical minutes
Operating profits or losses (€)
DRG = Diagnosis Related Groups; P-DRG = paediatric DRG
2009
Bed occupancy rate the same as in 2009, same mix of cases, no
P-DRG subventions
79.70 %
73 090
145 787.38░░
2009 100 %
2009 model with 100 % bed occupancy rate
100.00 %
91 700
1 124 999.15░░
A
No births, no P-DRGs
100.00 %
80 611
− 1 739 546.56▓▓
B
All pregnancies or children who are only slightly pathological or
not pathological at all are turned away
100.00 %
142 465
− 769 041.28▓▓
C
Only less profitable DRGs carried out
100.00 %
34 364
− 638 621.25▓▓
D
Surgical capacity (176 880 minutes) distributed across all
surgical DRGs; distribution of DRGs according to the share of
the respective DRG in the surgical minutes recorded for 2009; no
other DRGs; number of beds could be reduced by 6
84.42 %
176 880
− 503 067.48▓▓
E
Equal distribution of all bed capacities used in 2 009 for
deliveries across the 6 different O01 DRGs (c-sections); bed
occupancy rates for remaining DRGs remain the same
100.00 %
135 656
− 75 948.71▓▓
F
No P-DRGs
100.00 %
91 700
− 36 074.87▓▓
G
Only prenatal DRGs O64A/B and O65A/B and caesarean sections; the
same proportion of P-DRGs as in 2009
100.00 %
144 122
319 984.76▓▓
H
Surgical capacity (176 880 minutes) shared out completely and
evenly (16.67 %) across the 6 different c-section DRGs; no other
DRGs; 20 additional beds required which increases fixed
costs
155.90 %
176 880
555 085.64▓▓
I
Only non-surgical DRGs
100.00 %
0
1 170 873.41▒▒
J
Cf. models E and N; instead of the O01/O02 DRGs, this model
includes O60 DRGs (vaginal deliveries)
100.00 %
13 092
1 204 986.82▒▒
K
Cf. model G; instead of the caesarean sections, this model
includes O60 DRGs (vaginal deliveries)
100.00 %
0
1 214 574.84▒▒
L
All pregnancies and children with pathologies turned away
100.00 %
72 956
1 415 465.45▒▒
M
Share of bed occupancy rates of profitable DRGs are doubled,
share of less profitable DRGs are halved; adjustment of
calculated proportion of 100 % bed occupancy rate; same
proportion of P-DRGs as in 2009
100.00 %
96 038
1 742 922.43▒▒
N
Cf. models E and J; instead of O01 and O60 DRGs, this model
includes O02 DRGs (vaginal deliveries with complicating OR
procedure)
100.00 %
162 377
3 133 122.76▒▒
O
Proportion per DRG of the capacity constraint “inpatient days”
redistributed across other DRGs as profitably as possible
100.00 %
194 296
3 221 534.33▒▒
P
cf. model C, instead of the less profitable and loss-making DRGs
this model includes only the most profitable DRGs
100.00 %
92 918
5 108 788.48▒▒
Q
24 OR hours in the delivery room per day (525 600 surgical
minutes per year); only caesarean sections and P-DRGs; total of
63 additional beds required
275.01 %
525 600
9 532 718.69▒▒
The models “2009” and “2009 100 %” were the reference models (highlighted in light
grey in [Table 2 ]). The model 2009 calculated the
operating profits and losses, using the same bed occupancy rates and share of bed
occupancy rates per DRG as in 2009. The cost data and revenues on which the model
is
based and the potential factors which influence the basic model could have been
presented and discussed elsewhere [12 ].
The model 2009 100 % was particularly important as it served as a baseline with which
to compare other theoretical models with 100 % bed occupancy rates. If the operating
result of a theoretical model with a 100 % occupancy rate of available beds was
shown to be higher (highlighted in middle grey in [Table
2 ]) than the operating result calculated for the model 2009 100 %, this
model would recommend itself to the hospital management from an economic standpoint
(without taking account of the medical, ethical and legal issues involved). However,
if the operating result of a theoretical model was much below that calculated for
2009 100 % (highlighted in dark grey in [Table 2 ]), the
range of services provided would need to be reviewed and steps would need to be
taken to amend the range of services provided, if possible.
After adjusting the fixed costs proportionately to take account of the over- and
under-utilisation of the capacity constraint “inpatient days”, an operating profit
amounting to € 108 665.02 was calculated for model D, an operating loss of €
− 1 484 022.70 for model H and an operating profit of € 3 109 527.44 for model
Q.
Based on the operating profit of € 1 742 922.43 generated by doubling the share of
profitable DRGs and halving the proportion of less profitable DRGs, the case mix of
model M was best. The redistribution of the percentages of each DRG of the 2009 bed
occupancy rates of the UPF to other DRGs according to profitability in Model O led
to an operating profit of € 3 221 534.33.
As the simulations showed that the capacity constraint “inpatient days” was the most
significant factor, it was important to calculate how much operating results
depended on bed occupancy rates. It was assumed that the fixed costs remained the
same for all bed occupancy rates, whether bed occupancy rates were 100 % or not.
This meant that if services and treatments remained the same, there was a linear
relation between bed occupancy rates and operating results ([Table 3 ]). If the bed occupancy rate was only 50 %, this generated an
operating loss for the UPF of € − 1 285 437.66. A bed occupancy rate of 76.66 %
constituted the threshold between profit and loss with an operating profit of € 0.
If the bed occupancy rate was 79.70 % (as recorded for 2009) the UPF achieved an
operating profit of € 145 787.38. However, this figure does not factor in certain
costs, e.g. that 2 medical specialists are permanently on call to ensure 24 hour
readiness including coverage during vacation times and other downtimes, the impact
of interdisciplinary case conferences, the synergies achieved through the shared
services of surgical nurses, gynaecologists and obstetricians, etc. [12 ].
Table 3 Changes in operating results depending on bed
occupancy rates if the fixed costs remain the same.
Total bed occupancy rates (%)
Total length of stay in hospital (days)
Total surgical time (minutes)
CM I at the DRG level (€)
CM II at the DRG level (€)
Operating result for the UPF (CM III) (€)
CM = contribution margin; DRG = Diagnosis Related Groups
50 %
6 570
45 850
3 172 758.60
− 497 636.40
− 1 285 437.66
70 %
9 198
64 190
4 441 417.45
771 022.45
− 321 602.28
71 %
9 330
65 107
4 504 665.14
834 270.14
− 273 553.65
75 %
9 855
68 775
4 759 137.90
1 088 742.90
− 80 214.04
76.66 %
10 074
70 299
4 864 791.76
1 194 396.76
57.54
79.70 %
10 474
73 090
5 056 630.76
1 386 235.76
145 787.38
85 %
11 169
77 945
5 393 837.81
1 723 442.81
401 989.91
95 %
12 483
87 115
6 028 537.73
2 358 142.73
884 193.86
100 %
13 140
91 700
6 345 503.71
2 675 108.71
1 124 999.15
105 %
13 797
96 285
6 663 237.65
2 992 842.65
1 366 397.82
120 %
15 768
110 041
7 614 176.04
3 943 781.04
2 088 844.95
In general, a change in bed occupancy rates of one percentage point resulted in a
change in operating results of about € 48 000. This meant that if the UPF had a bed
occupancy rate of 100 %, the operating profit for the UPF would be € 1 124 999.15.
If the UPF was able to utilise beds in other hospital wards of the Gynaecology
Department of Erlangen University Hospital and thus achieve a bed occupancy rate of
120 % without being saddled with the additional fixed costs, it would be possible
to
achieve a profit of € 2 088 844.95 while offering the same services and treating the
same case mix. Such a capacity overload of bed occupancy rates could also be
achieved by discharging a patient and admitting another patient on the same day for
the same bed. The direct relationship between operating results, based on the real
model, and bed occupancy rates is shown in [Fig. 1 ].
Fig. 1 Relation between operating results and bed occupancy rates (PNC =
perinatal centre; CM = contribution margin).
Discussion
Given the financial problems the German healthcare system faces with the constantly
increasing pressures to reform and adapt, the agreements on minimum case numbers and
on the quality of structures, processes and outcomes, the issue how the services can
be financed is becoming ever more important. Perinatal centres must comply with the
criteria determined by the Joint Federal Committee which categorises the centres
into 4 different levels of care (levels I–IV). Because of the high demands placed
on
them and the high levels of staffing and financial resources required, level I
perinatal centres are under particularly high financial pressure. It is therefore
important that the management can distinguish between profitable and loss-making
DRGs and, where possible, take measures accordingly [13 ].
The fact that even obstetric departments with an inadequate infrastructure and
consistently low numbers of births still continue to operate would appear to
indicate that obstetric services are basically always profitable. However, a more
careful analysis shows that only certain services are financially viable in
obstetrics, and that the refinancing of invasive obstetric interventions within the
DRG system does not always cover costs [14 ]. Local
political considerations often also play a role by providing municipal subventions
for loss-making obstetric facilities to ensure their continued existence as this can
be the only way in certain districts of ensuring that women can give birth
locally.
The theoretical models developed for this study are based on real figures. They show
that even slight changes in bed occupancy rates and in the types of services offered
can quickly result in considerable losses, even in previously profitable level I
perinatal centres. It is therefore important to consider the impact of some factors
which hospitals cannot control, for example the patientʼs choice of hospital.
Particularly in obstetrics, recommendations made by the patientsʼ gynaecologist or
general practitioner are less important, and word-of-mouth comments and
recommendations by friends and acquaintances play a far more significant role
(“labour room tourism”) [15 ], [16 ].
The theoretical models were developed as a means of examining several issues in more
detail. First of all, it was shown that the number of cases treated and the bed
occupancy rate can have an enormous impact on revenues. In the real model, the UPF
had a bed occupancy rate of 79.70 %, a figure very close to the threshold of 77 %,
the point at which operating profits tipped over into operating losses. In practice,
this means that high numbers of cases must be viewed as very important for perinatal
centres. But the range of services provided also had an enormous impact. In the
different models the operating results calculated for bed occupancy rates of 100 %
ranged from operating losses of € − 1 739 546.56 to operating profits of €
5 108 788.48. This clearly shows how important the range of services is for costs
and revenues. The fact, demonstrated in this paper, that under the current DRG
system level I perinatal centres are only financially viable under certain
circumstances was already pointed out in the analysis by Seelbach-Göbel [13 ] which only analysed DRGs which ended with the birth of
the baby. The level I perinatal centre was only able to achieve an operating profit
because 46.95 % of births were DRG O60D, 16.37 % were DRG O60C and 4.89 % were DRG
O60B. These were not caesarean sections but vaginal deliveries without
complications. They were the only obstetric services where the figure calculated by
InEK (Institute for System of Hospital Remuneration) was higher than the actual
costs incurred by the perinatal centre. DRG O01A, which generates by far the highest
losses, amounting only to 0.30 % and therefore only played a marginal role in the
overall number of births recorded in the perinatal centre.
This study also selected and combined those services which would most increase the
hospitalʼs operating profits. A number of services currently offered were scrapped
and the contributions of certain services to full capacity were redistributed to
achieve the most profitable mix. Model P, for example, increased the operating
profit from € 1 124 999.15 to € 5 108 788.48 at a bed occupancy rate of 100 %.
Model A showed that the operating profits of perinatal centres depend very much on
the number of births. The 2009 case mix of the UPF with a bed occupancy rate of
100 % resulted in an operating profit of € 1 124 999.15. Model A used the same case
mix and bed occupancy rate but excluded births or P-DRGs and generated a loss of
€ − 1 739 546.56. This model did not take into account that children are born in
every perinatal centre and that certain minimum numbers of births are required for
every level I perinatal centre. But the simulation clearly showed that DRGs which
do
not immediately end with the birth of the child do not adequately cover the average
costs incurred by a level I perinatal centre. The lack of funding for perinatal
centres illustrated in model A could constitute a serious financial threat for
German obstetric clinics if the number of births continues to drop and obstetrical
developments in Germany follow the same path as in the Netherlands, where 30 % of
births in 2004 were home births [17 ].
Because of the higher relative weight of c-sections and the resulting higher revenues
it has occasionally been suggested that perinatal centres can only cover their costs
if they increase the rates of caesarean sections [18 ], [19 ], [20 ].
Heer et al. raised concerns about whether elective caesarean sections covered their
own costs in addition to those of vaginal deliveries [21 ]. The suspicion of Hornemann et al. [18 ] that
vaginal deliveries were being subsidised by primary caesarean sections was based on
the fact that their study did not take the high stand-by costs in obstetrics into
account. Only personnel costs which could be directly attributed to the respective
mode of delivery included in their analysis. As also pointed out by Schwenzer et al.
[22 ], even staffing costs which arise without a
specific birth event must be allocated to the births. Seelbach-Göbel [13 ] also commented that disregarding the costs of
non-medical infrastructure distorted the picture, as these costs generally amounted
to 25–30 % of the total costs of a perinatal centre.
All over the world the numbers of caesarean sections have been increasing for several
years. Germany is in the higher midrange with a rate of 28.5 % [23 ], [24 ]. One reason cited for the increased
rates is that they allow patients and physicians to plan and control the birth
better. Another potential reason discussed in the literature is the questionable
economic benefit compared to spontaneous deliveries. An analysis by the French
Association of Hospitals published in 2008 showed that c-section rates were higher
in private hospitals than in public hospitals, for example university hospitals,
although the latter provide care for more high-risk pregnancies, which could be
expected to lead to higher rates of caesarean sections [25 ]. According to Feige [20 ], only 2–3 % of
caesarean sections are so-called “sections on demand”, done at the request of the
patients themselves. Based on these figures and on the fact that caesarean sections
are medically indicated in only 15–20 % of cases, he deduced that around 10–15 % of
caesarean sections are done at the request of physicians.
Based on our calculations, if all relevant DRGs were compared directly, increasing
the rate of caesarean sections brought no financial benefits. We based this
assessment on the average contribution margin I per DRG and inpatient day, in other
words, on the remaining sum which covers the fixed costs after all variable costs
attributable to a DRG per inpatient day have been subtracted. The average figure for
all caesarean sections (DRGs O01A–F) was € 332.13, irrespective of the respective
number of cases, while the average figure for all vaginal births (DRGs O02A/B,
O60A–D) was € 426.56. After subjecting these figures to detailed analysis we found
that the difference remained even when the relevant DRGs were compared directly.
With an average contribution margin I per DRG case and inpatient day of € 448.28 and
given the capacity constraint “inpatient days”, caesarean sections for neonates with
a gestational age of more than 33 weeks and without complicating diagnoses (DRG
O01F) were more profitable than vaginal births without complicating diagnoses
(DRG O60D), which had a contribution margin of € 416.56. In comparison, caesarean
sections for premature babies until the end of the 25th week of gestation (DRG O01A:
€ 193) were economically less profitable than vaginal deliveries with complicating
OR procedures until the end of the 33rd week of gestation (DRG O02A: € 535.39) or
vaginal deliveries with at least one serious complicated diagnosis until the end of
the 33rd gestation week (DRG O60A: € 454.18). With an average contribution margin
I
per DRG case per inpatient day of € 448.28, DRG O01F (caesarean section without
complicating diagnosis > 33 weeks of gestation) is, from an obstetric point of
view, the only c-section DRG which is more profitable than the individual DRGs for
vaginal deliveries.
The economic disadvantages of c-section DRGs compared to vaginal births are shown
in
model E. In this model, all bed occupancy rates which were used in the real model
for birth-related DRGs (83.76 %) were evenly distributed (13.96 %) across the 6
different c-section DRGs, and all other percentages of total inpatient hospital days
of the remaining non birth-related DRGs compared to the calculation for 2009 were
left unchanged. If a bed occupancy rate of 100 % was assumed in model E, this
resulted in an operating loss of € − 75 948.71, which was lower than the operating
result of model J by € 1 280 935.53. In model J, all bed occupancy rates for
birth-related DRGs were evenly redistributed (20.94 %) across vaginal births DRGs
(DRGs O60A–D). If exclusively vaginal deliveries with complicating OR procedures
were posited, with a respective share of 41.88 % in the capacity constraint
“inpatient days” (DRG O02A/B), then the operating profit could even be increased to
€ 3 133 122.76 as shown in model N.
Schwenzer et al. [22 ] also came to the conclusion that in
Germany caesarean sections brought no economic benefit compared to vaginal
deliveries. Focussing on elective c-sections was only economically profitable if a
birth clinic or perinatal centre carried out c-sections only at certain times, as
this reduced or eliminated on-call costs [22 ]. Despite
the appeal of the additional revenues generated by c-sections it is important to
remember the higher costs involved. Moreover, it must be remembered that if the rate
of c-sections is increased, this will increase the costs of the remaining vaginal
births as the on-call costs have to be distributed across fewer vaginal births.
Seelbach-Göbel [13 ] showed that a level I perinatal centre
covers its actual costs predominantly by the revenues generated by spontaneous
births (DRG O60D) and would have an operating loss without them. The actual costs
of
vaginal births for the DRGs O60B (€ 123), O60C (€ 194) and O60D (€ 246) are below
the target costs in the InEK calculation. In contrast, the actual costs of
c-sections were far from covered by the InEK calculation. The shortfalls are shown
below with operating losses characterised by a negative sign. The costs of premature
births (O01A: € − 2992, O01B: € − 626), births with several complicating diagnoses
(O01C: € − 846, O01D: € − 343) and c-sections after the end of the 33rd week of
gestation (O01E: € − 332, O01F: € − 233) were not covered. This means that in a
level I perinatal centre the costs of complicated vaginal births before the 33rd
gestational week (O02A: € − 398, O60A: € − 1059) and the costs of vaginal deliveries
with complicating OR procedures after the 33rd week of gestation (O02B: € − 96)
could not be covered according to the costs calculated by InEK [13 ]. However, in contrast to our study, Seelbach-Göbel only included DRGs
which ended with the birth of the child. Thus, the remaining costs of in-hospital
stays during pregnancy were not included. When the real costs of university
hospitals were compared with the InEK calculation, only the costs of DRG O60D were
covered. All other birth-related DRGs in university hospitals resulted in quite
substantial losses [13 ].
An analysis of published reports for other countries comparing the costs of the
different types of births came to interesting results [22 ]. Henderson and colleagues [26 ] published a
review article where they summarised and analysed 49 international papers with
regard to the costs of vaginal deliveries without complications and c-sections. The
studies significantly showed that, on average, c-sections were more expensive than
vaginal births. The cost of a spontaneous birth without complications was between
£
629 and £ 1298, while the costs of c-sections were significantly higher, ranging
from £ 1238 to £ 3551.
In their analysis, Kazandjian et al. [27 ] came to the
conclusion that c-sections are not always more expensive than vaginal deliveries.
They viewed the perinatal centre as an interdependent economic unit and included the
costs of treating neonates in addition to birth mode-related costs. The average cost
for mother and child after a vaginal delivery, including the costs of the neonatal
intensive care unit, was thus $ 17 624.38 while the costs after a c-section were
only $ 13 805.47. If the costs of the neonatal intensive care unit were not
included, the cost of a c-section was $ 7529 and the cost of a vaginal delivery was
$ 5012.80.
Seelbach-Göbel [13 ] showed that in the current DRG system
complicated cases are not adequately remunerated. This issue, which has far-reaching
consequences for level I perinatal centres, was analysed in the theoretical models
B
and L. The models showed that from a purely economic standpoint it was better to
offer care to as few high-risk pregnancies as possible and instead to provide
services for pregnancies without associated pathologies.
Because of the current discussion about the minimum number of cases perinatal centres
need to treat when caring for premature babies and very low birth weight neonates,
it should be noted that some theoretical models are not easily compatible with
maintaining the centreʼs status as a level I perinatal centre. This particularly
applies to models A and L, in which either no births were possible or all
pathological pregnancies were excluded. Maintaining the current status of the UPF
was also questionable in the theoretical models C, D, F and H; in these models
enough extremely low birth weight neonates were born but they were not treated;
instead they were transferred to other hospitals. In contrast, if the case mixes of
the models I and K are used, the required case numbers of very low birth weight
neonates will not be achieved.
Based on our results, it would make more sense for certain hospitals to pursue a
classification as a level III or IV perinatal centre rather than as a level I or II.
But it must be emphasised that such reflections could jeopardise the comprehensive
coverage of high level perinatal centres. Offering additional incentives such as
supplementary payments or similar might be advisable from a health economic
perspective.
From a purely economic perspective, it might be expedient to downgrade a perinatal
centre as this would reduce labour costs. This aspect was not factored into the
theoretical models described here. The fact that changes to the range of services
and treatments provided would result in the partial or entire closure of certain
cost centres was also not included in our calculations. For the purposes of
simplification it was assumed that costs would be completely redistributed across
other cost centres. Thus, the theoretical models do not depict real scenarios; they
are theoretical marginal models which help increase the awareness of certain
economic aspects of healthcare. A good knowledge of marketing and marketing
strategies will be necessary if changes in services and treatments are implemented.
Lux et al. [16 ] showed how many multilayered and complex
the factors influencing patientsʼ choice of hospital are. If the management of a
hospital does not adapt its strategies and planning over the longer term to the
wishes of patients, operating profits could easily tip over into operating
losses.
The theoretical models described here are sure to prompt much discussion and they
have certain additional limitations. As the theoretical models are all based on the
results calculated for the real operating profit of the UPF in 2009, the limitations
and particularities of the UPF also apply to the theoretical models. It should also
be noted that the validity of some of the contribution margins I used here may
limited due to the low number of cases. Because of such variations, theoretical
models cannot provide a comprehensive financial picture. However, this is not the
aim of these theoretical models; the depiction of marginal cases contributes to
economic thinking and decision-making.
Because the paediatric department was not included in the calculations and because
obstetrics involve a very special type of patient, only limited conclusions on the
economic aspects of contribution margins can be drawn from this study. The question
arises whether examining the economic aspects of the care of premature babies,
neonates and their mothers and drawing conclusions based on economic considerations
alone is ethically defensible. Differentiating between “profitable” and “less
profitable” DRGs is not desirable, neither medically nor ethically, as this
differentiation focuses only on economic goals. In the theoretical models discussed
here, all ethical, medical and structural aspects and requirements were subordinated
to economic issues.
As level I perinatal centres are tertiary care centres, they are not permitted to
refuse any cases. When considering the range of services provided by a centre, it
should be remembered that in obstetrics the extent and type of treatment which may
be required cannot usually be planned ahead. Thus, when a patient is admitted to
hospital, it will not necessarily be immediately clear which DRG will be used for
the patient and her newborn baby at discharge.
Moreover it would also be necessary to study whether, when the bed occupancy rate
is
100 %, the fixed costs would also be higher than that cited in the calculation of
the UPFʼs operating result in 2009. In principle, fixed costs (such as personnel
costs) only increase if total capacity is increased. But in view of the fact that
employee resources are already spread quite thinly it is questionable whether
patients and their newborns could still be adequately cared for if the department
is
continually working to full capacity. In our calculations we therefore assumed that
the fixed costs would change depending on whether bed occupancy rates were higher
or
lower. To simplify matters in our calculations, it was assumed that changes in fixed
costs occurred progressively and not stepwise.
The validity of our theoretical models can currently not be verified against other
comparable models for German perinatal centres. Additional analysis, particularly
for perinatal centres providing other levels of care (levels II, III and IV), will
also be necessary.
