Journal of Pediatric Surgery (2011) 46, 617624

www.elsevier.com/locate/jpedsurg

Original articles

Costs of congenital diaphragmatic hernia repair in the

United Statesextracorporeal membrane oxygenation
foots the bill
Mehul V. Raval a,b,, Xue Wang a , Marleta Reynolds b , Anne C. Fischer c
a

Division of Research and Optimal Patient Care, American College of Surgeons, Chicago, IL 60611, USA
Division of Pediatric Surgery, Department of Surgery, Feinberg School of Medicine, Northwestern University,
Children's Memorial Hospital, Chicago, IL 60614, USA
c
Division of Pediatric Surgery, Department of Surgery, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA
b

Received 23 June 2010; revised 20 September 2010; accepted 20 September 2010

Key words:
Congenital diaphragmatic
hernia;
Extracorporeal membrane
oxygenation;
Cost;
Resource utilization;
Pediatric Surgery;
Kids' Inpatient Database

Abstract
Background: Congenital diaphragmatic hernia (CDH) is the costliest noncardiac congenital defect.
Extracorporeal membrane oxygenation (ECMO) is a treatment strategy offered to those babies with
CDH who would not otherwise survive on conventional therapy. The primary objective of our study was
to identify the leading source of expenditures in CDH care.
Methods: All patients surviving CDH repair were identified in the Kids' Inpatient Database (KID) from
1997 to 2006, with costs converted to 2006 US dollars. Patients were categorized into groups based on
severity of disease for comparison including CDH repair only, prolonged ventilator dependence, and
ECMO use. Factors associated with greater expenditures in CDH management were analyzed using a
regression model.
Results: Eight hundred thirty-nine patients from 213 hospitals were studied. Extracorporeal membrane
oxygenation use decreased from 18.2% in 1997 to 11.4% in 2006 (P = .002). Congenital diaphragmatic
hernia survivors managed with ECMO cost more than 2.4 times as much as CDH survivors requiring
only prolonged ventilation postrepair and 3.5 times as much as those with CDH repair only (both P b
.001). Age, multiplicity of diagnoses, patient transfer, inhaled nitric oxide use, prolonged ventilation,
and ECMO use were all associated with higher costs. Extracorporeal membrane oxygenation use was
the single most important factor associated with higher costs, increasing expenditures 2.4-fold (95%
confidence interval, 2.1-2.8). Though the CDH repair with ECMO group constituted 12.2% of patients,
this group has the highest median costs ($156,499.90/patient) and constitutes 28.5% of national costs
based on CDH survivors in the KID. Annual national cost for CDH survivors is $158 million based on
the KID, and projected burden for all CDH patients exceeds $250 million/year.
Conclusions: Extracorporeal membrane oxygenation use is the largest contributing factor to the
economic burden in CDH. With limited health care resources, judicious resource utilization in CDH care
merits further study.
2011 Elsevier Inc. All rights reserved.

Corresponding author. 633 N. Saint Clair St., 22nd Floor, Chicago, IL 60611-3211. Tel.: +1 312 202 5489; fax: +1 312 202 5062.
E-mail address: m-raval@md.northwestern.edu (M.V. Raval).
0022-3468/$ see front matter 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jpedsurg.2010.09.047

618
The first congenital diaphragmatic hernia (CDH) was
repaired by Dr Robert E. Gross in 1946 [1]. Since that
time, babies born with CDH remain one of the most
challenging populations to manage. Current reviews quote
optimistic overall survival rates of 80%, but these survival
rates have been hard fought and incorporated a variety of
advances in medications, respiratory technology, and
intensive care management [2,3]. Along with rapid
medical advancement, the costs of caring for children
with CDH have risen dramatically. Though exact dollar
figures are difficult to identify, single-center estimates
based on limited patients quote $250,000 per CDH case
and an annual national cost exceeding $230,000,000
making CDH the costliest noncardiac birth defect [4]. To
our knowledge, no studies have been conducted to provide
an objective large-scale national assessment of CDH costs
specifically targeting factors associated with increased
expenditure. It has long been suspected that extracorporeal
membrane oxygenation (ECMO) is a key component of
resource utilization in CDH management [5-7]. Thirty to
fifty percent of patients undergoing CDH repair are
estimated to require ECMO [6,8], yet the utility of
ECMO in CDH patients has been questioned [7,9,10]. In
turn, ECMO usage has been decreasing nationally, and
favorable outcomes have been reported with the use of
permissive hypercapnia and hypoventilation strategies [1012]. As the economic burden of health care in the United
States grows, objective assessments of judicious resource
utilization are needed.
The primary objective of this study was to analyze which
factors are the leading sources of expenditures in CDH
survivors. The secondary objective was to quantify the
economic burden of ECMO in CDH survivors after taking
into account patient-, hospital-, and disease-specific variables.

M.V. Raval et al.

data with Centers for Medicare and Medicaid Services
data. This conversion factor is hospital specific and
provides an estimate of all-payer inpatient costs [16]. A
linear predictive model was used to generate charge to cost
ratio estimates for hospitals missing provided conversion
factors. The KID also provides a ratio to weight the
charges represented by the database in terms of national
estimates. The weights are based on charge data from
2000, and subsequent weights were extrapolated using a
unique linear predictive model to all cases missing
provided national weights.

1.2. Patient selection

Patients with CDH were identified in the KID using
the International Classification of Diseases, Ninth
Revision, Clinical Modifications (ICD-9-CM) primary
diagnosis code 756.6 [17]. To obtain accurate cost
estimates, several inclusion and exclusion criteria guided
patient selection. Only patients who underwent CDH
repair were included based on specific ICD-9-CM
procedure codes (Table 1). Considering patients who
died may have had abbreviated hospitalizations and
therefore lower costs; only patients who survived to
discharge were included [18]. To minimize missing costs
accrued before patient transfer to the hospital of definitive
management, only patients who were born or arrived at
the treating facility within the first 24 hours of life were
selected. Patients missing charge data and age data were
excluded from analysis.

Table 1

Procedure codes used to identify CDH repair

ICD-9-CM
Procedure description
procedure codes

1. Methods

53.7

1.1. Data source

53.71

Data on patient encounters were analyzed from the

Agency for Healthcare Research and Qualitysponsored
Healthcare Cost and Utilization Project (HCUP) Kids'
Inpatient Database (KID) in 1997, 2000, 2003, and 2006
[13]. The KID is an administrative data set of patients aged
20 years or younger and contains data on more than 10
million hospitalizations from 38 states. The KID uses a
sampling of pediatric discharges and various outcomes of
interest, including total charges, and has been used to
compare freestanding children's hospitals to other institutions [14]. Charges from all years were converted to 2006
US dollars based on US Bureau of Labor Statistics
Consumer Price Index adjustment [15]. The KID also
provides a conversion factor for translating total charges
into actual costs using a system that has validated HCUP

53.72
53.75
53.8
53.8

53.81
53.82
53.83
53.84

Repair of diaphragmatic hernia, abdominal

approach
Laparoscopic repair of diaphragmatic hernia,
abdominal approach
Other and open repair of diaphragmatic
hernia, abdominal approach
Repair of diaphragmatic hernia, abdominal
approach, not otherwise specified
Repair of diaphragmatic hernia, thoracic
approach
Repair of diaphragmatic hernia with thoracic
approach, not otherwise specified;
thoracoabdominal repair of diaphragmatic
hernia
Plication of the diaphragm
Repair of parasternal hernia
Laparoscopic repair of diaphragmatic hernia,
with thoracic approach
Other and open repair of diaphragmatic
hernia, with thoracic approach

Costs of congenital diaphragmatic hernia repair

619

1.3. Variables of interest

2. Results

Various patient and hospital level variables were

compared. Patient factors studied included age at surgery,
sex, race, number of discharge diagnoses, insurance type,
and birth at treating hospital. The KID provides the age in
days at admission as well as the number of days from
admission to procedure. Using these values, age at surgery
was calculated. Age at surgery was divided into quartiles: 0
days, 1 to 2 days, 3 to 6 days, and 7 or more days. Race was
divided into white, black, Hispanic, and other. Patients had a
median of 9 discharge diagnoses recorded. The number of
discharge diagnoses was categorized into 5 or less, 6 to 10,
and 11 or more. Insurance type was categorized as Medicaid/
Medicare, private insurance, or other (including self-pay).
Hospital level factors included inborn or transfer status,
children's hospital designation, teaching status, urban or rural
location, and geographic regions. The KID assigns each hospital
to 1 of 4 center types based on National Association of
Children's Hospitals and Related Institutions designation:
freestanding children's hospital, children's specialty hospital,
children's units within general hospitals, or general/non
children's hospitals [19]. All children's hospitals were compared with general/nonchildren's hospitals. Geographic
regions were assigned based on hospital state location according
to US Census Bureau definitions [20]. Disease-specific factors
studied included diagnoses of prematurity or congenital cardiac
disease and use of inhaled nitric oxide, ECMO, or prolonged
ventilation. Prolonged ventilator dependence was identified
using ICD-9-CM procedure code 96.72 defined as continuous
invasive mechanical ventilation for 96 consecutive hours or
more. Extracorporeal membrane oxygenation use was identified
using ICD-9-CM procedure code 39.65.

From the KID, we identified 839 patients with CDH

who underwent repair in years 1997, 2000, 2003, and
2006. Table 2 compares selected variables of interest
compared by clinical group. Most patients (73.5%) in the
ECMO group had concurrent diagnostic codes for
prolonged ventilation. Patients who required ECMO were
operated on later than those who were managed with CDH
repair only or CDH repair with prolonged ventilation (P b
.001). Overall, 59.4% of patients were male, and 41.6% of
patients were white. Almost all patients were insured with
either Medicaid or private insurance. Whereas 27.7% of
patients had between 1 and 5 discharge diagnoses, 54.6%
had between 6 and 10 discharge diagnoses, and 30.4% had
11 or more discharge diagnoses. Patients that required
ECMO or prolonged ventilation also had a higher number
of associated discharge diagnoses (P b .001). One third of
patients were born at the hospital providing definitive
treatment, and two thirds were transferred from other
hospitals. A higher proportion of patients requiring ECMO

1.4. Statistical analysis

Patients were partitioned into 3 clinical groups: (1) CDH
repair only, (2) CDH repair and prolonged ventilator
dependence without ECMO use, and (3) CDH repair with
ECMO use with or without prolonged ventilation. Preliminary inferential statistics were calculated for all variables
using 2 tests for categorical data, Student t test to compare
means, and Wilcoxon rank sum test to compare medians.
Analysis of variance and Kruskal-Wallis 1-way analysis of
variance by ranks tests was used to compare the mean and
median costs accrued by the 3 clinical groups of interest.
Linear regression models were used to identify variables
associated with higher costs. Because of the right skewness of
the cost, a logarithmic transformation was used to serve the
restriction of the linear regression model. Coefficients were
determined for variables of interest with 95% confidence
intervals and translated to dollar figures. For all analyses,
statistical significance was set at P b .05. All P values are 2tailed. Statistical analyses were performed using SPSS,
version 18 (SPSS Corporation, Chicago, IL).

Table 2

Demographics of CDH repair by treatment groups

No. of patients (%)

CDH CDH repair

repair and
only prolonged
ventilation

CDH Overall
repair P value
and
ECMO

263
474
(31.3) (56.5)

102
(12.2)

Patient factors
Mean (SD) age
3.2
4.5
at surgery (d)
(7.2) (7.5)
Female (%)
40.7 39.7
Race (%)
White
39.2 41.6
Other
60.8 58.4
Insurance (%)
Medicaid
33.5 34.6
Other
66.5 65.4
Disease-related factors
No. of diagnoses (%)
1-10
93.3 80.0
11
6.7 20.0
Hospital factors
Hospital transfer (%)
63.1 66.2
Children's hospital (%) 74.1 78.3
Region (% by row)
Northeast
36.2 51.2
Midwest
24.1 61.0
South
38.2 47.3
West
27.4 66.7
Missing
35.8 46.0

9.3
(12.5)
45.1

b.001
.598
.420

48.0
52.0
.910
35.3
64.7
b.001
69.6
30.4
78.4
81.4

.019
.257
b.001

12.6
14.9
14.5
6.0
18.2

Note that some factors have been purposefully omitted and several
groups are collapsed for presentation, as no data involving 10 or fewer
observations may be published using HCUP KID data.

620

Fig. 1 Trends over time comparing CDH treatment groups

(overall P b .001).

were transferred (P = .02). The prolonged ventilation

patients represented the highest proportion requiring nitric
oxide treatment at 5.5% of patients (P b .001). There were
213 hospitals contributing between 1 to 34 cases, with a
median of 2 cases and mean of 4 cases per hospital.
Overall, 90% of hospitals were teaching centers, 96.8%
were located in urban settings, and 77.4% were designated
as children's hospitals. Compared with other regions where
CDH with ECMO use ranged from 12.6% to 14.5%, the
West region had the lowest ECMO use at 6.0% (P b .001).
Fig. 1 displays the various clinical groups compared by
year of study. The overall trend demonstrates a decrease in
the proportion of CDH patients managed with ECMO from
18.2% in 1997 to 11.4% in 2006, with a corresponding
rise in the number of CDH patients with prolonged
ventilation without ECMO use from 46.0% in 1997 to
64.4% in 2006 (overall P = .002).

2.1. Cost comparison including national estimates

Fig. 2 displays comparison of costs by clinical groupings.
Congenital diaphragmatic hernia repair and ECMO use cost
more than 3.5 times as much as CDH repair only and 2.4
times as much as CDH repair with prolonged ventilation

M.V. Raval et al.

(both P b .001). Exact cost figures are provided in Table 3.
As noted, the KID provides estimates that allow national
financial estimates to be determined using sampling
weights of patients and cases captured by the data set.
Based on these conversions, the national cost estimates for
CDH survivors are $158,696,578.65. Although the CDH
repair with ECMO group constitutes the smallest proportion of patients (12.2%), this group disproportionately has
the highest median costs per patient at $156,499.90 and
accounts for $45,276,956.37 (28.5%) of national costs for
CDH survivors. Based on estimates that 80% of patients
with CDH survive and assuming the burden of managing
patients that die resembles those who live, the national burden
for all CDH patients is projected as $198,370,723.20. If
patients who die more closely resemble those receiving
aggressive treatment, the national burden is projected to
$251,802,868.80. Based on a 60% overall survival for CDH
patients, the national estimates range from $264,494,297.80
to $406,980,091.10.

2.2. Factors associated with higher costs

Table 4 displays factors associated with higher costs.
After taking patient-, hospital-, regional-, and diseasespecific variables including inhaled nitric oxide use,
prolonged ventilation, and number of discharge diagnoses
into consideration, ECMO use was the single most
important factor associated with costs and increased costs
by 2.4-fold (95% confidence interval, 2.08-2.80). Adjusted
estimation using the linear regression model also identified
age at surgery of 7 days or older, greater number of
discharge diagnoses, patient transfer, inhaled nitric oxide
use, ECMO use, and prolonged ventilation were all
positively associated with costs. Patient sex, race, insurance status, care at a children's hospital, prematurity, or
the presence of cardiac anomalies were not associated with
costs. Care in the south was associated with lower costs
even when adjusted comparably.

3. Discussion

Fig. 2 Mean costs of CDH repair by treatment groups (overall

P b .001).

Patients with CDH represent one of the most complex and

costly populations. There has been a steady change in
management strategies without a clear consensus of those
best suited for survival. There has also been increasing
attention to judicious heath care resource utilization and the
controversies surrounding heroic life-saving interventions
such as ECMO [5]. Our study verifies that CDH patients
requiring ECMO constitute the highest costs and a
disproportionate amount of the national economic burden.
After taking patient-, hospital-, and disease-specific variables
into consideration, ECMO was the significant predictor of
cost. Compared with patients who did not use ECMO, CDH
patients managed with ECMO had a 240% increase in costs.

Costs of congenital diaphragmatic hernia repair

Table 3

621

Costs of CDH repair with national estimates by treatment groups

Total costs ($)

CDH repair only
CDH repair and prolonged ventilation
CDH repair and ECMO

No.

Mean

SD

Median

25th
percentile

75th
percentile

National cost
estimate

839
263
474
102

85,596.27
53,831.23
80,080.39
193,132.86

60,140.35
60,107.48
114,709.62
80,606.63

60,621.29
31,830.24
63,725.30
156,499.90

31,576.49
19,546.45
39,982.75
104,841.77

106,728.76
64,748.09
99,980.80
251,398.05

158,696,578.65
30,517,214.85
82,902,407.43
45,276,956.37

All significant to P b .001, with means being compared using analysis of variance and Bonferroni correction and medians being compared using KruskalWallis Test.

In 1995, Metkus et al [4] provided cost per survivor for 35

neonates with CDH managed from 1990 to 1993 at a single,
tertiary pediatric referral center. They estimated that the cost
per survivor was $98,000 in the non-ECMO group and
$365,000 in the ECMO group [4]. The 3.7-fold difference
noted was similar to the 3.5-fold difference we found using
the KID between CDH repair alone and CDH repair with
ECMO groups. The absolute cost figures and national
estimates used by Metkus et al [4] were based on a limited
sampling of itemized hospital bills (17 patients), reimbursement information (24 patients), and estimates of the
incidence of CDH and number of live births in 1991. In
contrast, our study uses administrative data from 813 patients
treated at 213 hospitals providing more generalizable cost
estimates. We have also incorporated verified weighted
sampling to normalize for regional and hospital effects. The
KID has been used to measure resource utilization in a
variety of rare, pediatric diseases, and the charge to cost
conversions, as well as methods for national estimation, has
been validated [14,16,21]. Using these conversions and
estimates, our study provides a more contemporary national
assessment of costs for CDH survivors as of 2006. We
estimate that the annual cost of caring for CDH repair
survivors reaches $160 million. This figure underestimates
the total economic burden of managing all patients with
CDH on several fronts. First, only survivors are considered
in this study. If current estimates of CDH survival are
accepted as 80%, our estimate brings the national burden
close to $200 million. These types of extrapolations must be
interpreted with caution, as the true costs of managing
patients with CDH are highly variable. Costs are low for
those who die within the first few hours of life but can also be
high for those who die after weeks of heroic treatment
including ECMO and long neonatal intensive care unit stays.
If most patients who die followed the latter group, estimated
national costs would exceed $250 million. These figures are
almost identical to estimates provided by Metkus et al [4]. If
survival estimates are closer to 60%, these estimates ranges
are $264 to $400 million. A second underestimation inherent
to our study is that only the initial costs of managing patients
with CDH are considered. None of these estimates account
for hidden costs from those diagnosed in utero and the costs
of associated with prenatal care and intervention. Furthermore, some CDH survivors have been shown to have

developmental delay and comorbidities that will ultimately

accrue life-long expenses that are not measurable using a
single administrative data set such as the KID [22].
Our study found that age at surgery of 7 days or older,
greater number of discharge diagnoses, patient transfer,
inhaled nitric oxide use, ECMO use, and prolonged
ventilation were all associated with higher costs. Sola et al
[23] recently queried the KID for overall resource utilization
in CDH and found that charges were higher based on race,
insurance, and number of diagnoses. Sola et al [23] provided
a low estimate of charges ($39,255 per patient) and included
patients based on a diagnosis of CDH but did not address
definitive surgical repair of the CDH, CDH survival, or
include specific information on the use of ECMO. Our study
did not find a correlation between race or insurance status
and higher costs. This difference may be partially explained
by our use of cost data, reflecting actual costs, as compared
with charge data, which is highly variable and hospital
specific. Metkus et al [4] identified birth weight and the need
for ECMO as the only independent variables associated with
high costs in their single-institution report. They did not find
a correlation between higher costs and gestational age, patch
closure, and number of prenatal diagnoses [4]. Stevens et al
[6] retrospectively reviewed 2165 term infants spanning
1976 to 2001 from the Extracorporeal Life Support
Organization registry and found that greater birth weight,
absence of a prenatal diagnosis, higher 5-minute Apgar
score, and earlier year of birth were independent predictors of
survival for CDH patients on ECMO, but they did not
provide data on costs.
Extracorporeal membrane oxygenation is a resourceintensive treatment with wide variation in use depending on
center philosophy ranging from liberal application by some
and considered as rescue therapy by others. Although there
have been no prospective cost analyses studies for ECMO in
the pediatric population, there have been 2 retrospective
reviews. Van Litsenburg [24] et al estimated ECMO
utilization for a variety of respiratory and cardiac indications
at $176,467. Vats et al [25] gauged the charges for 20
patients placed on ECMO for a variety of indications at
nearly $200,000 and estimated that the cost of ECMO spread
over a lifetime at $4190 per year of life saved. These
estimates resemble the difference in mean costs of
$139,301.63 between the CDH-only group and the CDH

622

M.V. Raval et al.

Table 4

Factors associated with costs for CDH repair

Linear regression

95.0% confidence interval

Adjusted estimates

Lower bound

Upper bound

30,312.43

24,018.33

38,255.93

Reference group
0.726
1.013
1.282

22,005.29
30,691.81
38,859.34

15,109.02
21,071.19
26,333.32

32,049.27
44,704.99
57,343.64

b.001
.865
.002

Reference group
0.999

30,286.58

21,892.63

41,898.92

.985

Reference group
0.889
1.041
1.103
1.007

26,943.02
31,562.80
33,446.28
30,517.10

16,443.95
21,579.60
22,113.75
21,489.01

44,145.52
46,164.45
50,586.34
43,338.13

.376
.591
.286
.911

Reference group
1.411
1.906

42,780.60
57,787.80

30,259.63
39,069.98

60,482.57
85,473.05

b.001
b.001

Reference group
1.064
1.138

32,238.16
34,505.14

23,088.49
22,544.46

45,013.73
52,811.40

.232
.188

Reference group
1.119

33,915.92

24,213.57

47,506.00

.035

Reference group
1.063

32,216.86

22,753.21

45,616.70

.299

Reference group
1.063
0.749
1.150
0.772

32,222.24
22,711.23
34,866.75
23,389.86

21,436.41
15,118.52
23,867.45
15,839.18

48,435.03
34,117.12
50,935.07
34,540.03

.493
.001
.061
.001

Reference group
1.367

41,426.12

25,777.72

66,573.93

.011

Reference group
2.414

73,180.52

49,941.48

107,233.30

b.001

Reference group
1.256

38,079.60

27,276.13

53,162.11

b.001

Reference group
1.136

34,423.33

22,957.27

51,616.16

.148

Reference group
0.925

28,050.21

19,740.45

39,857.98

.200

Coefficient
Overall ($)
Age at surgery (d)
0
1-2
3-6
7
Sex
Male
Female
Race
White
Black
Hispanic
Other
Missing
No. of diagnoses
1-5
6-10
11
Insurance
Medicaid
Private
Other
Transfer status
Inborn
Transfer
Hospital type
Non-children's
Children's
Region
Northeast
Midwest
South
West
Missing
Nitric oxide use
No
Yes
ECMO use
No
Yes
Prolonged ventilation
No
Yes
Prematurity
No
Yes
Cardiac anomalies
No
Yes

Presented coefficients have been exponentiated.

with ECMO groups in our study. Ultimately, the costeffectiveness of ECMO use specifically for CDH is an area
of discussion, and there is a lack of prospective data guiding

the debate especially in light of the evolving management of

CDH [5,7,9,25]. As a point of reference, data from the KID
have been used to study charges for congenital heart disease,

Costs of congenital diaphragmatic hernia repair

and the 90th percentile cut point defining high-resource users
was $192,272 per patient, with some patients exceeding $1
million in charges [26]. When studying congenital heart
disease patients who require ventricular assist devices, mean
hospital costs are even higher at $624,798 per patient [27].
Trends show a decrease in the use of ECMO for patients
with CDH. Comparing 1997 with 2006, we found a 6.8%
decrease in CDH patients with ECMO and a 18.4% increase
in the number of CDH patients managed with prolonged
ventilation and no ECMO. Similar findings have been
reported from the Extracorporeal Life Support Organization
registry as well as from the Congenital Diaphragmatic Hernia
Study Group [6,10]. Lally et al followed patients with
diaphragmatic agenesis from 20 hospitals and found a
decrease in ECMO use, an increase in inhaled nitric oxide
use, and a trend toward improved survival when comparing
patients from 1995 to 1996 to patients from 2003 to 2004
[10]. These observations support the notion that relatively
newer therapies such as high-frequency oscillatory ventilation, inhaled nitric oxide, and gentle ventilation may lead to
improved survival without the need for ECMO [11,28].
We also found that CDH repair at an earlier day of life
was associated with lower costs among survivors and may be
a marker for those cases that are less complex along the
physiologic spectrum of CDH patients. The optimal timing
of CDH repair remains uncertain with a lack of high-quality
randomized controlled trials and more than 20 observational,
nonrandomized or case series reports debating the merits and
limitations of early vs delayed repair [29-32]. The trend
toward delayed repair has led to an increase in the number of
CDH patients on ECMO before CDH repair, with some
patients being repaired while on ECMO [33,34].
Based on size and scope, the KID serves as large
repository of information and provides insight as to current
practice trends and changes over time. The KID is based on
administrative data centered on billing, and therefore, use of
the KID for studies evaluating national resource utilization is
one of its best uses. The limitations of our study are those
inherent in the use of an administrative data set that lacks
clinical measures specific to CDH and ECMO use. Inherent
are the limitations with hospital discharge data including
inaccuracies in coding and discordance between clinically
relevant information and information used primarily for
billing. As noted, there are several limitations to extrapolating the data available in the KID to all CDH cases,
specifically to patients who died and patients who were
transferred from one institution to another. When selecting
cases, we attempted to minimize information lost by costs
being split by 2 hospitals. Ultimately, patients who are
managed at a hospital, cannot be stabilized, and require
transfer may represent a sicker cohort that would incur
high expenditures. The obvious example is the transfer or a
CDH patient after several days of difficult management
specifically to a center offering ECMO services. Whereas
many studies quote the use of ECMO for managing CDH at
30% to 50% of patients, our analysis found that 12.2% of

623
those CDH patients who survived were managed with
ECMO. If anything, our cost figures have underestimated the
financial burden of CDH in the United States by limiting
estimates for patients who either were transferred or died.
Lastly, there is no doubt that CDH patients who require
ECMO represent a sicker cohort of patients. We attempted to
account for the severity of illness by accounting for the
number of discharge diagnoses and estimating the severity of
pulmonary disease using billing codes for ventilator
dependence. By using 3 clinical groups and not just
comparing ECMO to no-ECMO, we attempt to depict a
group of patients, those with prolonged ventilation without
ECMO, who may more closely resemble the severity of
illness represented in the ECMO group. These approximations do not fully capture severity of illness, which may be
better measured with additional data such as severity of
pulmonary hypertension, size of CDH defects, and other
more clinical factors. In light of these limitations, we hope
that our study provides those treating CDH with a
contemporary estimate of costs based on a large cohort of
nationally representative patients and may assist researchers
interested in studying the cost-effectiveness of various
treatment options for this difficult-to-treat population.

4. Conclusions
Using information from over a decade of cases captured
by the KID data set from more than 200 hospitals, we
estimate the national costs of caring for CDH survivors at
nearly $160 million dollars per year, extrapolated up to
$250 million for all CDH care. Extracorporeal membrane
oxygenation use in CDH survivors was associated with
tripling of costs and a disproportionate amount of the
national economic burden of caring for CDH patients. In
addition to ECMO, age at surgery of 7 days or older,
greater number of diagnoses, patient transfer, inhaled nitric
oxide use, and prolonged ventilation were associated with
higher costs. These figures and findings only comprise the
initial management costs as reported to an administrative
data set. The true costs to survivors of CDH repair are
likely to be much higher given the multiple diagnoses and
follow-up requirement. Ultimately, the true cost of ECMO
for CDH patients will need to be evaluated in a
prospective fashion. With limited health care resources
and economic constraints, utilization of ECMO in CDH
patients merits further study.

Acknowledgments
M.V.R. participates in the American College of Surgeons
Clinical Scholars in Residence Program. M.V.R. is
supported by the John Gray Research Fellowship and Aida
and Daniel L. Rice.

624

References
[1] Gross RE. Congenital hernia of the diaphragm. Am J Dis Child
1946;71:579-92.
[2] Harting MT, Lally KP. Surgical management of neonates with congenital
diaphragmatic hernia. Semin Pediatr Surg 2007;16(2):109-14.
[3] Ladd WE, Gross RE. Congenital diaphragmatic hernia. N Engl J Med
1940(223):917-25.
[4] Metkus AP, Esserman L, Sola A, et al. Cost per anomaly: what does a
diaphragmatic hernia cost? J Pediatr Surg 1995;30(2):226-30.
[5] Crow S, Fischer AC, Schears RM. Extracorporeal life support:
utilization, cost, controversy, and ethics of trying to save lives. Semin
Cardiothorac Vasc Anesth 2009;13(3):183-91.
[6] Stevens TP, Chess PR, McConnochie KM, et al. Survival in earlyand late-term infants with congenital diaphragmatic hernia treated
with extracorporeal membrane oxygenation. Pediatrics 2002;110(3):
590-6.
[7] Khan AM, Lally KP. The role of extracorporeal membrane
oxygenation in the management of infants with congenital diaphragmatic hernia. Semin Perinatol 2005;29(2):118-22.
[8] Sreenan C, Etches P, Osiovich H. The western Canadian experience
with congenital diaphragmatic hernia: perinatal factors predictive of
extracorporeal membrane oxygenation and death. Pediatr Surg Int
2001;17(2-3):196-200.
[9] Keshen TH, Gursoy M, Shew SB, et al. Does extracorporeal
membrane oxygenation benefit neonates with congenital diaphragmatic hernia? Application of a predictive equation. J Pediatr Surg
1997;32(6):818-22.
[10] Lally KP, Lally PA, Van Meurs KP, et al. Treatment evolution in highrisk congenital diaphragmatic hernia: ten years' experience with
diaphragmatic agenesis. Ann Surg 2006;244(4):505-13.
[11] Boloker J, Bateman DA, Wung JT, et al. Congenital diaphragmatic
hernia in 120 infants treated consecutively with permissive hypercapnia/spontaneous respiration/elective repair. J Pediatr Surg 2002;37(3):
357-66.
[12] Logan JW, Rice HE, Goldberg RN, et al. Congenital diaphragmatic
hernia: a systematic review and summary of best-evidence practice
strategies. J Perinatol 2007;27(9):535-49.
[13] HCUP Kids' Inpatient Database (KID). Healthcare Cost and
Utilization Project (HCUP). 1997, 2000, 2003, and 2006. Agency
for Healthcare Research and Quality, Rockville, MD. Available at
www.hcup-us.ahrq.gov/kidoverview.jsp.
[14] Merenstein D, Egleston B, Diener-West M. Lengths of stay and
costs associated with children's hospitals. Pediatrics 2005;115(4):
839-44.
[15] Williamson SH. Six ways to compute the relative value of a U.S. dollar
amount, 1790 to present; 2009. Available at: http://www.measuring
worth.com/uscompare/.
[16] Healthcare Costs and Utilization Project - Cost-to-Charge Ratio Files.
Agency for Healthcare Research and Quality, Rockville, MD.
Available at http://www.hcup-us.ahrq.gov/db/state/costtocharge.jsp.

M.V. Raval et al.

[17] International Classification of Diseases-9-CM 2008. Chicago (Ill):
American Medical Association; 2007.
[18] Rogowski J. Using economic information in a quality improvement
collaborative. Pediatrics 2003;111(4 Pt 2):e411-8.
[19] NACHRINational Association of Children's Hospitals and Related
Institutions. Available at http://www.childrenshospitals.net/ Vol.
January 19, 2010, 2010.
[20] U.S. Census Bureau; 2000. Available at http://www.census.gov/main/
www/cen2000.html.
[21] Odetola FO, Gebremariam A, Freed GL. Patient and hospital correlates
of clinical outcomes and resource utilization in severe pediatric sepsis.
Pediatrics 2007;119(3):487-94.
[22] Davis PJ, Firmin RK, Manktelow B, et al. Long-term outcome following
extracorporeal membrane oxygenation for congenital diaphragmatic
hernia: the UK experience. J Pediatr 2004;144(3):309-15.
[23] Sola JE, Cheung MC, Bronson SN, et al. Resource utilization for
congenital diaphragmatic hernia in the United States (abstract only).
J Surg Res 2010;158(2):282-4.
[24] Van Litsenburg R, De Mos N, Edgell D, et al. Resource use and health
outcomes of paediatric extracorporeal membrane oxygenation. Arch
Dis Child Fetal Neonatal Ed 2005;90(2):F176-7.
[25] Vats A, Pettignano R, Culler S, et al. Cost of extracorporeal life
support in pediatric patients with acute respiratory failure. Crit Care
Med 1998;26(9):1587-92.
[26] Connor JA, Gauvreau K, Jenkins KJ. Factors associated with increased
resource utilization for congenital heart disease. Pediatrics 2005;116
(3):689-95.
[27] Mahle WT, Ianucci G, Vincent RN, et al. Costs associated with
ventricular assist device use in children. Ann Thorac Surg 2008;86(5):
1592-7.
[28] Desfrere L, Jarreau PH, Dommergues M, et al. Impact of delayed
repair and elective high-frequency oscillatory ventilation on survival of
antenatally diagnosed congenital diaphragmatic hernia: first application of these strategies in the more severe subgroup of antenatally
diagnosed newborns. Intensive Care Med 2000;26(7):934-41.
[29] de la Hunt MN, Madden N, Scott JE, et al. Is delayed surgery really
better for congenital diaphragmatic hernia? A prospective randomized
clinical trial. J Pediatr Surg 1996;31(11):1554-6.
[30] Nio M, Haase G, Kennaugh J, et al. A prospective randomized trial of
delayed versus immediate repair of congenital diaphragmatic hernia.
J Pediatr Surg 1994;29(5):618-21.
[31] Moyer V, Moya F, Tibboel R, et al. Late versus early surgical
correction for congenital diaphragmatic hernia in newborn infants.
Cochrane Database Syst Rev 2002(3):CD001695.
[32] West KW, Bengston K, Rescorla FJ, et al. Delayed surgical repair and
ECMO improves survival in congenital diaphragmatic hernia. Ann
Surg 1992;216(4):454-60 discussion 460-2.
[33] Lally KP, Paranka MS, Roden J, et al. Congenital diaphragmatic hernia.
Stabilization and repair on ECMO. Ann Surg 1992;216(5):569-73.
[34] Dassinger MS, Copeland DR, Gossett J, et al. Early repair of
congenital diaphragmatic hernia on extracorporeal membrane oxygenation. J Pediatr Surg 2010;45(4):693-7.