Journal of Pediatric Surgery 53 (2018) 52–59

Contents lists available at ScienceDirect

Journal of Pediatric Surgery

journal homepage: www.elsevier.com/locate/jpedsurg

Effective methods to decrease surgical site infections in pediatric

gastrointestinal surgery
Andrew B. Nordin a,b, Stephen P. Sales a, Gail E. Besner a,c, Marc A. Levitt a,c,d,
Richard J. Wood a,c,d, Brian D. Kenney a,c,⁎
a
Nationwide Children's Hospital, Department of Pediatric Surgery, Columbus, OH
b
State University of New York University at Buffalo, Department of General Surgery, Buffalo, NY
c
The Ohio State University College of Medicine, Columbus, OH
d
Nationwide Children's Hospital, Center for Colorectal and Pelvic Reconstruction, Columbus, OH

a r t i c l e i n f o a b s t r a c t

Article history: Background: Gastrointestinal (GI) surgeries represent a significant proportion of the surgical site infection (SSI)
Received 26 September 2017 burden in pediatric patients, resulting in significant morbidity. Previous studies have shown that perioperative
Accepted 5 October 2017 bundles reduce SSIs, but few have focused on pediatric GI operations. We hypothesized that a GI bundle would
decrease SSI rates, length of stay (LOS), and hospital charges.
Key words: Methods: After establishing baseline SSI rates, a GI bundle was created and implemented in November 2014. We
Surgical site infection
prospectively collected data including demographics, procedure type, LOS, inpatient charges, bundle compliance,
GI
Bundle
and SSI development. We analyzed SSI rates, LOS, and charges using process control charts.
Results: The baseline SSI rate for all GI operations was 3.4%, which increased to 7.1%, then decreased to 4.7%. Mid-
gut/hindgut and stoma closure SSI rates decreased from 11.3% to 8.0% (p b 0.05) and 21.4% to 7.9%, respectively
(p b 0.05). Although overall LOS and charges were unchanged, average LOS for midgut/hindgut surgeries and
stoma closures decreased from 20.3 to 13.6 days (p = 0.015) and 12.6 to 7.9 days (p = 0.04), respectively.
Stoma closure charges decreased from $94,262 to $50,088 (p = 0.01).
Conclusions: Our perioperative GI bundle decreased SSI rates, primarily among midgut/hindgut operations. Bun-
dle usage decreased LOS and charges most effectively in stoma closures.
Type of study: Prognosis Study.
Level of evidence: Level 2.
© 2017 Elsevier Inc. All rights reserved.

Surgical site infections (SSIs) represent a common source of postop-
erative morbidity and mortality. Their development contributes to in-
creased length of stay (LOS), increased hospital costs, and decreased
quality of life [1–6]. Gastrointestinal (GI) operations comprise a signifi-
cant proportion of this overall disease burden [2]. Multiple risk factors
for SSI have been identified in adults including hypothermia, hypergly-
cemia, obesity, medical comorbidities, and wound classification; to the
extent that these factors can be prevented or minimized, SSI rates can
be reduced [7–10]. Appropriate preoperative skin preparation and anti-
biotic administration have also been shown to decrease the incidence of
SSI [11,12]. For colorectal operations, the usage of a bowel preparation
may also reduce infection rates. Adult studies suggest that the
⁎ Corresponding author at: Nationwide Children's Hospital, Department of Pediatric
Surgery, 700 Children's Drive, Columbus, OH, 43205. Tel.: +1 614 722 3819.
E-mail address: Brian.Kenney@nationwidechildrens.org (B.D. Kenney).
combination of a mechanical bowel prep and oral antibiotics decreases
SSIs, whereas a mechanical prep alone may have the opposite effect
[1,13,14]. For bowel prep, as for many other specific SSI-reducing prac-
tices, the majority of evidence comes from studies in adult patients. Al-
though pediatric surgeons often apply adult principles to their patients,
a dedicated pediatric focus is sorely needed.
Many of these practices can be combined into perioperative bundles,
which minimize variation in patient care. Several studies have demon-
strated that perioperative care bundles effectively decrease SSI rates in
adults undergoing colorectal surgery [15,16]. The body of literature sur-
rounding pediatric perioperative bundles is growing, and multicenter
studies have demonstrated that bundles reduce SSI rates in pediatric
spine, cardiac and neurosurgical procedures [17,18]. Additionally, rec-
ognizing that stoma closures typically have exceptionally high SSI
rates, a recent study showed that bundle usage reduced stoma closure
SSI rates in pediatric patients [19]. Since SSIs incur a significant medical

https://doi.org/10.1016/j.jpedsurg.2017.10.018
0022-3468/© 2017 Elsevier Inc. All rights reserved.
 A.B. Nordin et al. / Journal of Pediatric Surgery 53 (2018) 52–59 53

Table 1 1. Methods
GI surgery perioperative bundle.

Preop Bowel Prep 1.1. Patient data collection

• Bowel prep required for all patients undergoing a procedure involving the rec-
tum unless a proximal stoma is present and is not being concomitantly reversed
• Inpatient regimen
o GoLytely 25 mL/kg/h × 4 h
o Neomycin 15 mg/kg/dose (×3 doses)
o Erythromycin 20 mg/kg/dose (×3 doses)
▪ 10 mg/kg/dose for neonates b30 days old
Preop Cleansing
• Patients N2 months: clean the abdomen with 2% chlorhexidine gluconate wipes
• Patients b2 months: clean the abdomen with antimicrobial wipes
Preop Warming
• Measure patient temperature 1 h prior to operation
• Apply convection warming blanket for all patients with initial
temperature b 36.5 °C
o Recheck temperature every 30 min
Preop Antibiotics
• Administer appropriate antibiotic to finish within 60 min of incision
o Cefazolin for foregut and HPB procedures. Redose as needed
o Cefoxitin for midgut/hindgut procedures. Redose as needed
▪ Gentamicin/clindamycin for patients with penicillin allergies
▪ Ampicillin/gentamicin acceptable for neonates within first week of life; add
clindamycin after first week
o If patient is on adequate systemic antibiotics prior to the procedure, no addi-
tional antibiotics are needed. Redose as needed
Skin Prep
• Chlorhexidine for all patients N2 months or N1 kg
• 10% povidone-iodine for patients b2 months or b1 kg
Closing Protocol (for procedures in which the bowel has been opened and fascial
closure is needed)
• Prior to fascial closure:
o All staff change gloves
o Redrape the surgical field
o Remove all dirty instruments; use clean instruments for fascia and wound
closure
and financial burden, we developed a comprehensive perioperative GI
bundle to decrease SSI rates, and hypothesized that its use would in
turn result in decreased overall LOS and inpatient hospital costs.
We determined baseline 30-day SSI rates from January 2014 to No-
vember 2014 through a retrospective chart review of all patients under-
going GI surgeries at a tertiary care free-standing pediatric hospital. We
defined SSIs according to accepted Centers for Disease Control defini-
tions [20]. Since then, we have continued to prospectively review the
charts of all patients undergoing GI surgeries. Data points collected in-
clude age, sex, body mass index (BMI), ASA class, pre-operative location,
procedure performed, total hospital LOS, intensive care unit (ICU) LOS,
number of ICU admissions, total number of GI operations, and total
30-day inpatient charges. Procedures were broadly categorized as either
“Foregut,” “Hepatopancreaticobiliary (HPB),” or “Midgut/Hindgut,” and
also as elective, urgent, or emergent. Appendectomies and trauma oper-
ations were excluded.
1.2. Bundle development
Based on best practice recommendations from the adult literature,
we developed a perioperative bundle for all GI surgeries, which went
into effect in November 2014 (Table 1). The bundle was subsequently
modified in January 2016 to include closing protocol for all stoma clo-
sures. We monitored and analyzed bundle compliance on a monthly
basis beginning in August 2014, to establish baseline practice patterns.
We also provided ongoing feedback to improve compliance with all
bundle components.
1.3. Data analysis
SSI rates were calculated as the number of infections divided by the
total number of GI procedures for each month, and were tracked over
time using Shewhart process control charts. Separate SSI rates were cal-
culated for each procedure category, with special attention to midgut/
hindgut procedures and stoma closures, since these groups generally
have higher reported SSI rates [2,9]. With these charts, baseline and pro-
cess stage means are established based on an absolute minimum of 6
consecutive values, although process stages may extend longer provid-
ed that they reasonably explain variation among the contained data
points. Control limits are set at 3 standard deviations from the mean.
Variation within the control limits is considered acceptable, or common
cause variation, whereas points outside these limits represent special
cause variation and imply an external confounding factor. For each

Fig. 1. Overall SSI rates for all GI surgeries. Control limits are set at 3 standard deviations from the mean. Process means are established based on a minimum of 6 consecutive values, and
shifts in the process mean are significant at the p = 0.05 level.
54 A.B. Nordin et al. / Journal of Pediatric Surgery 53 (2018) 52–59

Table 2
Patient characteristics.

Without SSI With SSI p value

Overall
Sex (% female) 53.30% 66.00% 0.07
Age (years) 7.26 8.33 0.095
BMI 19.70 19.92 0.43
n = 891 ASA 2.33 2.28 0.27
Foregut
Sex (% female) 57.10% 100.00% 0.14
Age (years) 8.75 13.67 0.12
n = 199
BMI 22.18 33.45 0.07
ASA 2.67 3.00 0.18
HPB
Sex (% female) 46.70% 100.00% 0.35
Age (years) 13.72 14.50 0.41
BMI 25.25 31.71 0.17
n = 131 ASA 2.19 2.00 0.33
Midgut/Hindgut
Sex (% female) 66.70% 62.50% 0.04
Age (years) 6.40 7.73 0.07
BMI 17.33 18.58 0.024
n = 546 ASA 2.24 2.25 0.45
Stoma Closure
Sex (% female) 40.60% 61.10% 0.11
Age (years) 6.66 4.87 0.14
BMI 18.05 18.14 0.47
n = 119 ASA 2.22 2.33 0.21

shift in the process mean, we calculated significance using Student's t
test to a significance level of p b 0.05. We established goals for overall in-
fection rates and bundle compliance at the project's inception. These
were based on reasonably attainable target infection and compliance
rates, respectively, and were not based on a power analysis since this
project was designed as an ongoing quality improvement initiative
without a set enrollment number.
To demonstrate the effects of our bundle on decreasing SSI burden,
we also analyzed LOS and 30-day inpatient charges using control charts,
using the same methodology as for our SSI rate analysis. Since our pop-
ulation was heterogeneous in terms of age and specific surgical proce-
dures, we then performed a matched cohort analysis to validate the
impact of SSI development on our outcome measures. Each patient un-
dergoing an elective procedure who developed an SSI was matched
based on age, BMI, preoperative location (home, inpatient floor, ICU,
neonatal ICU), and procedure performed with two similar patients
who did not have an SSI. Demographic characteristics between patients
with SSIs and those without were compared using Student's t-test or
chi-square test as appropriate. Average hospital charges, LOS, ICU LOS
and number of GI operations were compared between groups using
Student's t-test, with significance levels set at p = 0.05. No Institutional
Review Board approval was required since this is an ongoing quality im-
provement project designed to improve patient care at our institution
by reducing SSI rates.
2. Results
From January 2014 to November 2014, prior to bundle implementa-
tion, there were an average of 55 GI operations performed per month.
Our initial overall SSI rate was 3.4%, which increased to 7.1% during
this time period (p = 0.05). In the 23 months following bundle imple-
mentation, there were a total of 1595 GI operations performed on
1474 patients, representing an increase of approximately 15 cases per
month. These include 1328 (83.2%) elective, 151 (9.5%) urgent and
105 (6.6%) emergent cases, with respective SSI rates of 6.6%, 6.0% and
2.9%. Of the elective cases, 891 (67.1%) were planned procedures, in

Fig. 2. Midgut/hindgut SSI rates. The decrease in the mean SSI rate from 11.3% to 8.0% is statistically significant (p = 0.03) and chronologically correlated with a significant increase in
bundle compliance.
 A.B. Nordin et al. / Journal of Pediatric Surgery 53 (2018) 52–59 55

Fig. 3. Stoma closure SSI rates. The decrease in the mean SSI rate from 21.4% to 7.9% is statistically significant (p = 0.02) and chronologically correlated with our addition of the closing
protocol for these operations.

which patients were admitted from home within 1–2 days for the pur-
poses of their operation. 120 cases (7.5%) were repeat GI operations per-
formed during the same index admission. Since we began using the
bundle, our overall SSI rate decreased to 4.7%, despite an increase in sur-
gical volume (p = 0.03; Fig. 1). Overall 30-day mortality following bun-
dle implementation was 1.15% (n = 17 patients), and only 0.11% (n = 1
patient) in planned elective cases. There was no significant difference in
BMI or ASA class between patients who developed an SSI and those that
did not. Patients with an SSI tended to be female and older, although
these differences were not statistically significant (Table 2).
Following bundle implementation, we began tracking SSI rates
by procedure category and found that midgut/hindgut and stoma
closure SSI rates were initially much higher than the overall SSI
rate. However, as bundle compliance increased, both significantly
decreased. Our SSI rate for midgut/hindgut procedures decreased
from 11.3% to 8.0% and remained stable at that rate for over
12 months (p = 0.03; Fig. 2). The baseline stoma closure SSI rate
was 21.4%, and this decreased to 7.9% following bundle implemen-
tation (p = 0.03; Fig. 3). Foregut and HPB SSI rates were 2.3% and
1.1% respectively, and did not significantly change following bun-
dle implementation (data not shown). Overall bundle compliance
increased from 43% to 80% since the bundle was introduced
(p b 0.001; Fig. 4). These rates represent fully compliant cases
only, in which all bundle elements were performed.
In our matched cohort analysis, we matched 53 patients who devel-
oped an SSI after an elective procedure with 106 similar patients who
did not develop an SSI. Among this cohort, there were no significant dif-
ferences in age, BMI or ASA class between patients that did and did not
develop an SSI, although a significantly higher number of female pa-
tients developed SSIs (data not shown). Patients who did not develop
an SSI had a shorter LOS (8.3 v 13.9 days; p = 0.002) and incurred
fewer charges ($80,997 v $131,897; p = 0.002) compared to those
who had an SSI (Table 3). Midgut/Hindgut procedures, including
stoma closures specifically, showed significant increases in the number
of additional GI operations, ICU LOS, ICU admissions and total average
charges associated with the presence of an SSI. Overall LOS was signifi-
cantly higher for midgut/hindgut procedures that developed SSIs, but
there was no statistical difference in LOS for stoma closure patients. In
this analysis, the presence of an SSI had no significant impact on out-
comes for HPB and foregut procedures.
To demonstrate the direct effects of our bundle on LOS and charges,
we also constructed control charts to track these outcomes. In our over-
all population, bundle application did not result in significant changes in
LOS (data not shown). However, average LOS in midgut/hindgut cases
decreased from 20.3 to 13.6 days (p = 0.02; Fig. 5) and from 12.6 to
7.9 days for stoma closures (p = 0.04; Fig. 6). Average 30-day inpatient
charges did not significantly change following bundle implementation
for the overall population (data not shown). Among midgut/hindgut

Fig. 4. Overall bundle compliance rates. Procedures must be compliant with each component of the bundle: if one component is not performed, the case is considered noncompliant.
Compliance increased from 43% to 80% (p b 0.001).
56 A.B. Nordin et al. / Journal of Pediatric Surgery 53 (2018) 52–59

cases, charges decreased over time, although not to a significant degree

p value

0.012
0.002
0.002
(data not shown). In contrast, average stoma closure charges decreased

NS
NS
from $94,262 to $50,088 (p = 0.01; Fig. 7).

$68,167.47 (64,440.88)
$83,683.18 (85,458.89)
$80,996.74 (82,997.61)

$33,296.99 (9291.87)
3. Discussion
Mean 30-day inpatient charges ($)
We have demonstrated that the application of a perioperative bun-

(141,162.20)
$140,921.90
Without SSI dle decreases SSIs for pediatric GI procedures, most notably among mid-
gut/hindgut operations and stoma closures specifically. Furthermore,
for these groups of procedures, bundle use is associated with decreased
LOS and a trend towards decreased hospital charges. To our knowledge,

$137,827.20(137,435.90)
this is the first report in the pediatric literature demonstrating the suc-
$28,753.67 (12,184.60) cessful application of a comprehensive GI bundle using quality improve-
ment processes.

(159,593.20)
(138,524.40)
(131,250.60)

Prior to any intervention, our overall SSI rate had initially increased
$141,628.75
$197,420.20
$131,896.75

from baseline in the prebundle period, which may represent a surveil-

p value With SSI

lance bias. Following bundle implementation, the overall SSI rate de-
creased, despite the fact that we established a dedicated colorectal
center during this timeframe, increasing both the number and complex-
0.0004
0.015
0.002
Mean no. of ICU admissions

ity of surgical procedures in a population already at increased risk for SSI

NS
NS

development. Our observed decrease in overall SSI rates was chronolog-

0.06 (0.24)
0.12 (0.32)
0.33 (0.60)
0.12 (0.32)

ically associated with both a decrease in midgut/hindgut SSI rates, and

0.0 (0.0)
Without

an increase in GI bundle compliance. These findings suggest that in-

creased bundle compliance decreased our total SSI rates, primarily
SSI

owing to decreases among midgut/hindgut surgeries. Stoma closure

SSI rates did not drop initially, but once we added a closing protocol
p value With SSI

0.0 (0.0)

(0.59)
(0.61)
(1.05)
(0.73)

for these procedures, they similarly decreased. There was no effect on

0.29

0.39
0.36

0.67

SSI rates for foregut or HPB procedures, which is likely a consequence

of their low baseline SSI rates.
0.022

0.031
0.011

Although we substantially reduced SSI rates for midgut/hindgut pro-

NS
NS

cedures and stoma closures in particular, our overall SSI rates were
Mean ICU LOS (days)

0.44 (1.71)

0.06 (0.24)
0.51 (1.99)

1.13 (2.99)

largely unchanged from the initial baseline rate. Given the relatively re-
0.0 (0.0)
Without

cent decrease in midgut/hindgut SSI rates and the month-to-month var-

iability, more time may need to elapse to observe a decrease in overall
SSI

SSI rates. Alternatively, the lack of significant change may be related to

b0.00001 1.98 (6.84)
b0.00001 2.21 (6.82)

2.56 (5.10)

bundle compliance, which is crucial to effecting change [17,18]. Al-

With SSI

0.0 (0.0)

(10.43)

though our compliance increased significantly from 43% to 80%, it may

3.61

need to increase further to cause our infection rates to decrease. Nota-

bly, our compliance control chart considers cases noncompliant if
even one aspect of the bundle is missed. Since the bundle involves mul-
p value

0.0004
0.029

tiple providers across different departments, providing ongoing feed-

Mean no. of GI surgeries

NS

back and reinforcing bundle components have been, and will be,
1.03 (0.17)

1.03 (0.13)
1.03 (0.17)

1.04 (0.20)

crucial to improving compliance.

1.00 (0.0)
Without

Extensive evidence exists in the adult literature to show that SSIs

prolong hospitalization and increase hospital costs, ICU admissions,
SSI

hospital readmissions, and overall mortality [21,22]. SSIs following GI

1.00 (0.0)

operations may prolong average LOS by up to 10 days, and average hos-

p value With SSI

(0.63)

(0.83)
(0.62)

(0.61)

pital costs increase accordingly: estimated cost savings for the preven-
1.38

1.56
1.36

1.31

tion of a single SSI range from $3500 to almost $30,000 [3,22–24]. We

confirm that for elective pediatric operations, the presence of an SSI
0.009
0.002

was associated with an increase in overall and ICU LOS, the number of
NS

NS
NS

ICU admissions, number of repeat GI operations, and average 30-day in-

Comparative outcomes from matched cohort analysis.

patient charges. These findings were most prominent among midgut/

9.67 (10.46)
Mean LOS (days)

Without SSI

0.50 (0.50)
8.31 (9.24)

hindgut procedures, particularly stoma closures, validating that this

(11.78)
(80.48)

10.69
43.20

group of patients benefits the most from our specific bundle measures.
It is insufficient, however, to merely demonstrate the adverse conse-
quences of SSI development, as these are well-documented and do not
inherently suggest that our bundle avoided such outcomes. For this rea-
1.00 (0.0)
With SSI

(12.61)

(10.92)
(12.39)

(59.84)

son, we constructed separate control charts to explicitly evaluate the

Midgut/Hindgut 14.44

14.50
All GI Operations 13.68

39.15

bundle's effect on LOS and charges. Bundle implementation did not re-
sult in a significant decrease in LOS or inpatient charges for the entire
Stoma closures

population of GI operations. However, as expected, decreases in SSIs in

midgut/hindgut and stoma closure operations translated to decreased
Foregut

LOS and, for stoma closures, a decrease in charges. Although we did

Table 3

HPB

not observe a significant reduction in charges among midgut/hindgut

procedures, there does appear to be a declining trend in charges with
 A.B. Nordin et al. / Journal of Pediatric Surgery 53 (2018) 52–59 57

Fig. 5. Average midgut/hindgut length of stay (LOS). Average LOS decreased from 20.3 to 13.6 days (p = 0.015), and our control limits narrowed, suggesting decreased variation among
cases.

Fig. 6. Average stoma closure length of stay. Average LOS decreased from 12.6 to 7.9 days (p = 0.04).

Fig. 7. Average stoma closure 30-day inpatient charges. Average charges decreased from $94,262 to $50,088 (p = 0.01). No significant decreases in inpatient charges were observed for
midgut/hindgut procedures or the overall group of GI surgeries, which may be because of variations in patients' clinical courses.
58 A.B. Nordin et al. / Journal of Pediatric Surgery 53 (2018) 52–59
time, and further data collection and analysis are required. Finally, it is
worth noting that, for each decrease in LOS, our control limits narrowed
correspondingly, suggesting that bundle usage reduced variation in our
perioperative care.
As in adults, SSIs in children following GI operations can have
substantial ramifications. In an analysis of 2011 Pediatric National
Surgery Quality Improvement Program (NSQIP) data, Saito et al. re-
ported an overall SSI rate of 2.3% for general pediatric surgery cases
[25]. An additional study of 2012–2013 pediatric NSQIP data showed
overall and colorectal SSI rates of 2% and 5.9%, respectively [2]. Al-
though our overall SSI rate was higher than these studies, we focused
specifically on GI surgeries, excluding many of the clean cases which
were included in the NSQIP data. Our overall mortality rate was min-
imal and consistent with the rates reported by NSQIP. Regarding the
financial implications of SSIs, Sparling et al. demonstrated a potential
savings of more than $27,000 in a matched cohort cost analysis of 16
pediatric patients undergoing clean or clean-contaminated cases
[24]. We have demonstrated that even greater estimated cost reduc-
tions may be attainable for elective GI surgeries, and most of this dif-
ference can be attributable to reducing SSIs in midgut/hindgut
procedures.
Although previous multicenter studies have demonstrated bundle
utility in pediatric cardiac, spine and ventriculoperitoneal shunt surger-
ies, only one other study evaluated GI bundle usage in children, and it
showed that application of a stoma closure-specific bundle decreased
SSI rates from 42.8% to 13.9% [17–19]. In the current study, we con-
firmed that a perioperative bundle is effective at decreasing SSI rates
for GI surgeries, particularly midgut/hindgut procedures and stoma clo-
sures. In these categories, bundle implementation was associated with
decreased average LOS and inpatient charges. Although we have made
significant progress in combating a major cause of postoperative mor-
bidity, we believe that additional improvements are still possible, and
we have therefore continued to evaluate different potential bundle
modifications to further reduce SSI rates.
This study is unique in the pediatric literature, but does contain cer-
tain limitations. First, as with all quality improvement work, we are un-
able to demonstrate that our bundle, or any specific bundle component,
directly caused a decrease in SSI rates. We do, however, demonstrate a
strong temporal correlation between increased bundle compliance
and decreased infection rates. Secondly, as our results reflect changing
trends over time, they are subject to confounding influences, including
variations in surgeon or trainee practice, patient populations, disease
complexity, and even our data collection processes. For instance, we
suspect that the increase in SSI rates during the prebundle period repre-
sents a surveillance bias, wherein we identified additional infections as
we began more closely monitoring our processes. Our reduction in SSI
rates may also have been influenced by the Hawthorne effect rather
than a direct result of the bundle elements: surgeons may have altered
their practice in response to bundle compliance monitoring. Further-
more, this work represents only cases performed at a single institution,
and our practices and findings may not be applicable to other centers.
Our large sample population should at least partially mitigate this
disadvantage.
Finally, our financial analysis was limited in that we used 30-day in-
patient charges only. These values likely overestimate actual costs, may
not be generalizable, and do not include outpatient charges for clinic
visits and wound care, particularly if provided outside of our healthcare
system. Our analysis did not incorporate the charges associated with
specific bundle elements, but the effects of excluding these charges
should be minimal, since all practices in the bundle are routinely per-
formed. We also did not account for the indirect and intangible costs
of SSIs to patients and their families, such as additional clinic visits,
time lost from work or school, or impact on quality of life. However, in-
cluding these costs would only have accentuated the substantial bur-
dens of developing an SSI, and our results may actually underestimate
the magnitude of the bundle's impact.
Our findings open multiple opportunities for further study, which in
turn could result in further modifications to our bundle and improve-
ments in SSI rates. Previous retrospective work in children identified
race, age b 30 days, postoperative ICU admission, and the presence of
a Foley catheter or other implantable device as significant SSI risk fac-
tors [5]. Our ongoing project may be able to validate these findings,
identify additional risk factors, and potentially intervene through bun-
dle modification. Bundle changes, such as requiring a closing protocol
for stoma closures, have already resulted in decreased SSI rates for this
subgroup, and additional adjustments to the bundle may yield further
benefit. For example, we currently have no standard practice for closing
the stoma site wound during stoma closures; however, research in adults
has demonstrated that a pursestring skin closure results in the lowest SSI
rates compared to other closure methods [26]. Future studies will also
evaluate the impact of SSIs and bundle application in neonates specifical-
ly, who have been shown to have higher SSI rates than the general pedi-
atric population [25,27]. Finally, our bundle should be prospectively
validated in a multicenter study. Applying our bundle at other institutions
may highlight areas for additional modification to further reduce SSI rates.
4. Conclusion
In conclusion, pediatric patients who developed an SSI had increased
overall LOS, number of ICU admissions and LOS, and inpatient hospital
charges. SSI rates were reduced through application of and compliance
with our GI bundle, particularly among midgut/hindgut infections and
stoma closures, and reduced infections in these groups translated to de-
creased LOS and, for stoma closures, decreased hospital charges. We
confirm that perioperative bundles, which are commonly used in adults,
result in similar benefits when applied to pediatric GI surgeries. Further
evaluation of this bundle in a multicenter study may result in greater
opportunities to significantly improve postoperative surgical care in
children.
Author contributions
Andrew Nordin was involved in data collection, analysis, and manu-
script drafting and revision. Stephen Sales assisted with data analysis
and interpretation, as well as manuscript revision. Gail Besner, Marc
Levitt and Richard Wood all contributed to data interpretation and crit-
ical manuscript revision. Brian Kenney was responsible for study design,
data interpretation and manuscript revision.
Appendix A. Discussions
Andrew Nordin, Columbus OH.
GRAHAM COSPER (Charlotte NC) Can you describe your closing
protocol?
ANDREW NORDIN Thank you. Our closing protocol involves changing
the surgeon's gloves prior to closing fascia for all open cases,
so that does not include any laparoscopic cases.
DOUGLAS BARNHART (Salt Lake City UT) A couple of comments. First
of all, thanks for doing this and presenting it. I think it is a
good example of where we have a problem with a high mor-
bidity burden, and hopefully your work will serve as a begin-
ning for a national collaborative to work on this problem.
I have a couple of specific questions. You did not break it
down into types, classification of surgical site infection in
terms of superficial, deep, or organ space. Did you look at
that, and did you see a difference? That is one question. The
other question I had just in terms of your data, did you mon-
itor for whether or not there was a change in leaving the
wounds open? Because if the change is in fact in superficial
surgical site infection, if you have gained increased attention
to this, and surgeons are leaving the wound open, by
 A.B. Nordin et al. / Journal of Pediatric Surgery 53 (2018) 52–59
definition they will not get a superficial surgical site infection.
Thanks again for your work.
ANDREW NORDIN To answer the first question regarding the type of
wound infections, we certainly have that data and have not
analyzed it specifically over time but have not seen a major
difference in the wound classification types. With regards to
the method of wound closure, that is something which is
sort of our next step in continuation of this project and some-
thing that we are working on analyzing.
KURT HEISS (Atlanta GA) Could you talk a little bit about bundle fail-
ure? Was it an act of civil disobedience, or was it team failure?
And then secondly, when you talk about using a back table
protocol, you know, a single site procedure with a 3-cm
wound is really an open procedure as well, right, and we
think about using that for single site colectomies and things
of that nature. Lastly, do you purse string your stoma closures
to leave it as semi open-type thing?
ANDREW NORDIN With regard to the method of closure, for the single
site procedures that is something that I think would still fall
under our laparoscopic domain and not necessarily require
the closing protocol. With regards to the ostomy closures and
the purse string method of closure, that is something that has
been demonstrated to reduce superficial infections in adults
and again specifically for ostomy closures that is an area in
which we are working to determine whether the method of
closure has any relationship to the development of an infection.
In regards to bundle compliance, the control chart that I
showed represents overall bundle compliance, and so that is
a composite of the preoperative staff doing the cleansing and
warming, the operating room staff, and the surgeon compli-
ance. I think in each of those areas and some more than others
there are ways in which we have room for improvement.
References
[1] Rangel SJ, Islam S, St Peter SD, et al. Prevention of infectious complications after
elective colorectal surgery in children: an American pediatric surgical association
outcomes and clinical trials committee comprehensive review. J Pediatr Surg
2015;50(1):192–200.
[2] Feng C, Sidhwa F, Cameron DB, et al. Rates and burden of surgical site infections
associated with pediatric colorectal surgery: insight from the National Surgery
Quality Improvement Program. J Pediatr Surg 2016;51(6):970–4.
[3] Keenan JE, Speicher PJ, Thacker JK, et al. The preventive surgical site infection bundle
in colorectal surgery: an effective approach to surgical site infection reduction and
health care cost savings. JAMA Surg 2014;149(10):1045–52.
[4] Azoury SC, Farrow NE, Hu QL, et al. Postoperative abdominal wound
infection—epidemiology, risk factors, identification and management. Chronic
Wound Care Manage Res 2015;2:137–48.
59
[5] Whitehouse JD, Friedman ND, Kirkland KB, et al. The impact of surgical-site infec-
tions following orthopedic surgery at a community hospital and a university hospi-
tal: adverse quality of life, excess length of stay, and extra cost. Infect Control Hosp
Epidemiol 2002;23(4):183–9.
[6] Anthony T, Long J, Hynan LS, et al. Surgical complications exert a lasting effect on
disease-specific health-related quality of life for patients with colorectal cancer. Sur-
gery 2003;134(2):119–25.
[7] Bucher BT, Guth RM, Elward AM, et al. Risk factors and outcomes of surgical site in-
fection in children. J Am Coll Surg 2011;212(6):1033–8.
[8] Tang R, Chen HH, Wang YL, et al. Risk factors for surgical site infection after elective
resection of the colon and Rectum: a single-center prospective study of 2,809 con-
secutive patients. Ann Surg 2001;234(2):181–9.
[9] Kim MS, Kim HK, Kim DY, et al. The influence of nutritional assessment on the out-
come of ostomy takedown. J Korean Soc Coloproctol 2012;28(3):145–51.
[10] Kurz A, Sessler DI, Lenhardt R. Perioperative normothermia to reduce the incidence
of surgical-wound infection and shorten hospitalization. N Engl J Med 1996;
334(19):1209–15.
[11] Lee I, Agarwal RK, Lee BY, et al. Systematic review and cost analysis comparing use of
chlorhexidine with use of iodine for preoperative skin antisepsis to prevent surgical
site infection. Infect Control Hosp Epidemiol 2010;31(12):1–21.
[12] Hawn MT, Gray SH, Vick CC, et al. Timely administration of prophylactic antibiotics
for major surgical procedures. J Am Coll Surg 2006;203(6):803–11.
[13] Leys CM, Austin MT, Pietsch JB, et al. Elective intestinal operations in infants and
children without mechanical bowel preparation: a pilot study. J Pediatr Surg 2005;
40(6):978–82.
[14] Serrurier K, Liu J, Breckler F, et al. A multicenter evaluation of the role of mechanical
bowel preparation in pediatric colostomy takedown. J Pediatr Surg 2012;47(1):190–3.
[15] Tanner J, Padley W, Assadian O, et al. Do surgical care bundles reduce the risk of sur-
gical site infections in patients undergoing colorectal surgery? A systemic review
and cohort meta-analysis of 8,515 patients. Surgery 2015;158(1):66–77.
[16] Waits SA, Fritze D, Banerjee M, et al. Developing an argument for bundled interven-
tions to reduce surgical site infection in colorectal surgery. Surgery 2014;155(4):
602–6.
[17] Toltzis P, O'Riordan M, Cunningham DJ, et al. A statewide collaborative to reduce pe-
diatric surgical site infections. Pediatrics 2014;134(4):e1–7.
[18] Schaffzin JK, Harte L, Marquette S, et al. Surgical site infection reduction by the solu-
tions for patient safety hospital engagement network. Pediatrics 2015;136(5):
e1353–60.
[19] Porras-Hernandez J, Bracho-Blanchet E, Tovilla-Mercado J, et al. A standardized peri-
operative surgical site infection care process among children with stoma closure: a
before–after study. World J Surg 2008;32:2316–23.
[20] Mangram AJ, Horan TC, Pearson ML, et al. Guideline for prevention of surgical site
infection, 1999. Hospital infection control practices advisory committee. Am J Infect
Control 1999;27:97–134.
[21] Gagliardi AR, Eskicioglu C, McKenzie M, et al. Identifying opportunities for quality
improvement in surgical site infection prevention. Am J Infect Control 2009;37(5):
398–402.
[22] de Lissovoy G, Fraeman K, Hutchins V, et al. Surgical site infection: incidence and im-
pact on hospital utilization and treatment costs. Am J Infect Control 2009;37(5):
387–97.
[23] Broex ECJ, van Asselt ADI, Bruggeman CA, et al. Surgical site infections: how high are
the costs? J Hosp Infect 2009;72(3):193–201.
[24] Sparling KW, Ryckman FR, Schoettker PJ, et al. Financial impact of failing to prevent
surgical site infections. Qual Manag Health Care 2007;16(3):219–25.
[25] Saito JM, Chen LE, Hall BL, et al. Risk-adjusted hospital outcomes for children's sur-
gery. Pediatrics 2013;132(3):e677–88.
[26] Hsieh MC, Kuo LT, Chi CC, et al. Pursestring closure versus conventional primary clo-
sure following stomal reversal to reduce surgical site infection rate: a meta-analysis
of randomized controlled trials. Dis Colon Rectum 2015;58(8):808–15.
[27] Segal I, Kang C, Albersheim SG, et al. Surgical site infections in infants admitted to
the neonatal intensive care unit. J Pediatr Surg 2014;49(3):381–4.