A Quality Improvement Initiative to

Reduce Hospitalizations for Low-risk

Diabetic Ketoacidosis
Kelly R. Bergmann, DO,a M. Jennifer Abuzzahab, MD,b Joe Arms, MD,a Gretchen Cutler, PhD,a Heidi Vander Velden, MS,a
Ted Simper, PharmD,c Eric Christensen, PhD,d Dave Watson, PhD,e Anupam Kharbanda, MDa

BACKGROUND AND OBJECTIVES: Children

with established type 1 diabetes (T1D) who abstract
present to the emergency department (ED) with mild diabetic ketoacidosis
(DKA) are often hospitalized, although outpatient management may be
appropriate. Our aim was to reduce hospitalization rates for children with a
Departments of Emergency Medicine, cPharmacy,
established T1D presenting to our ED with mild DKA who were considered e
Research and Sponsored Programs, and bPediatric
low risk for progression of illness. Endocrinology and McNeely Diabetes Center, Children’s
Minnesota, St Paul, Minnesota; and dCollege of Continuing
METHODS: We conducted a quality improvement initiative between January 1, and Professional Studies, University of Minnesota,
2012, and December 31, 2018 among children and young adults #21 years of Minneapolis, Minnesota

age with established T1D presenting to our tertiary care ED with low-risk Dr Bergmann conceptualized and designed the study,
DKA. Children transferred to our institution were excluded. DKA severity was developed the database, conducted the medical
record review, conducted the data analysis,
classified as low, medium, or high risk on the basis of laboratory and clinical participated in the interpretation of the results,
criteria. Our quality improvement initiative consisted of development and drafted the manuscript as written, and critically
implementation of an evidence-based treatment guideline after review by reviewed and revised the manuscript; Drs
Abuzzahab, Arms, Kharbanda, and Simper aided in
a multidisciplinary team. Our primary outcome was hospitalization rate, and the study concept and design, participated in the
our balancing measure was 3-day ED revisits. Statistical process control interpretation of the results, and critically reviewed
methods were used to evaluate outcome changes. and revised the manuscript; Drs Cutler, Christensen,
and Watson aided in the study concept and design,
RESULTS: Weidentified 165 patients presenting with low-risk DKA. The baseline performed the data analysis, participated in the
preimplementation hospitalization rate was 74% (95% confidence interval interpretation of the results, and critically reviewed
64%–82%), and after implementation, this decreased to 55% (95% and revised the manuscript; Ms Vander Velden
conducted the medical record review, aided in the
confidence interval 42%–67%) (219%; P = .011). The postimplementation development of the database, participated in the
hospitalization rate revealed special cause variation. One patient in the interpretation of the results, and critically reviewed
postimplementation period returned to the ED within 3 days but did not have and revised the manuscript; and all authors
approved the final manuscript as submitted and
DKA and was not hospitalized. agree to be accountable for all aspects of the work.
Hospitalization rates for children and young adults presenting to
CONCLUSIONS: DOI: https://doi.org/10.1542/peds.2019-1104
the ED with low-risk DKA can be safely reduced without an increase in ED Accepted for publication Nov 20, 2019
revisits. Address correspondence to Kelly R. Bergmann, DO,
Department of Emergency Medicine, Children’s
Minnesota, 2525 Chicago Ave South, Mail Stop 32-
1488, Minneapolis, MN 55404. E-mail:
Approximately 208 000 children in the that the mean cost of hospital kelly.bergmann@childrensmn.org
United States have type 1 diabetes admission for DKA is ∼$7140.5 Much of PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online,
(T1D),1 and the incidence is increasing this cost is attributable to inpatient 1098-4275).
by 1.8% annually.1,2 Up to 30% of care, including treatment with Copyright © 2020 by the American Academy of
children present with diabetic intravenous (IV) insulin and admission Pediatrics
ketoacidosis (DKA) at initial diagnosis to an ICU.5,7
of T1D,3,4 and 20% to 47% have at least To cite: Bergmann KR, Abuzzahab MJ, Arms J,
1 subsequent readmission for DKA,5,6 Current treatment recommendations et al. A Quality Improvement Initiative to Reduce
Hospitalizations for Low-risk Diabetic Ketoacidosis.
resulting in substantial medical allow for variability in the care of
Pediatrics. 2020;145(3):e20191104
expenditure. Recent literature suggests children with DKA.8,9 As a result, there

Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020

PEDIATRICS Volume 145, number 3, March 2020:e20191104 QUALITY REPORT
remains uncertainty regarding
optimal treatment strategies,
particularly for those presenting with
mild DKA, defined as follows: (1)
venous pH of 7.2 to 7.3 or
a bicarbonate level of 10 to 15 mmol/
L; (2) hyperglycemia, with blood
sugar levels . 200 mg/dL; and (3)
ketonemia or ketonuria.8 This is
supported by recent literature
suggesting that resource use and
readmission rates for DKA vary
significantly across US children’s
hospitals.5 Previous research has
revealed that most children with mild
DKA have resolution of acidosis
within hours of initiating
treatment.10,11 More recent studies
suggest that patients with DKA who
are managed with subcutaneous
insulin, which may be appropriate for
children with mild DKA in whom
discharge is anticipated, have similar
time to recovery of acidosis compared
with those treated with IV
insulin.7,12–15 Children who present
to the emergency department (ED)
with mild acidosis, established T1D,
knowledge of sick-day management
and ability to perform home care, and
no social conditions or comorbid
illness that would impede discharge
may be considered low risk and
managed at home with
outpatient care.
Preliminary data from our institution
revealed that 74% of children with
established T1D presenting to our ED
with low-risk DKA were hospitalized.
We hypothesized that we could safely
reduce admission rates in this
population. Therefore, our aim for the
quality improvement (QI) initiative
was to reduce hospitalization rates by
10% for children with established
T1D presenting to our ED with low-
risk DKA over a 2-year period.
METHODS
Study Design and Setting
We conducted a QI initiative among
children and young adults #21 years
of age with DKA who presented to 2
Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020
2 BERGMANN et al
urban, tertiary care pediatric EDs.
The QI initiative was implemented
April 1, 2016, and we examined
patients from January 1, 2012, to
December 31, 2018 (pre- and
postimplementation). Both EDs are
part of a single pediatric health
system, staffed by the same pediatric
emergency medicine physicians and
subspecialist services, and have
a combined 95 000 visits annually.
The study was approved by our
institutional review board.
Study Population and DKA Risk
Categorization
We identified children with DKA via
a combination of an electronic
medical record (EMR) query and
manual chart review. First, we
developed an automated monthly
query of our EMR database to identify
all children with an ED discharge
diagnosis for DKA (International
Classification of Diseases, Ninth
Revision and International
Classification of Diseases, 10th
Revision codes 250.11 and 250.13 and
E10.10 and E10.11, respectively).
Second, we performed a manual chart
review to confirm the presence of
DKA among children with T1D.
DKA was defined as having all of the
following: (1) venous pH , 7.3 or
a bicarbonate level , 15 mmol/L, (2)
urine or serum sample positive for
ketones, and (3) a glucose level .
200 mg/dL. Although definitions of
DKA severity may vary slightly,
classically, DKA severity has been
defined solely by laboratory criteria,
in which mild is pH , 7.30 or
a bicarbonate level , 15 mmol/L,
moderate is pH , 7.20 or
a bicarbonate level , 10 mmol/L, and
severe is pH , 7.10 or a bicarbonate
level , 5 mmol/L.8 However, given
that factors other than the degree of
acidosis may influence the true
severity of DKA, related
complications, and, in particular, the
decision to hospitalize a child, we
further classified children with DKA
as follows: low risk (pH . 7.2 or
bicarbonate level . 10 mmol/L,
established T1D, reliable home cares
as determined by the treating
provider after discussion with the
family or patient, or insulin pump
failure with ability to correct),
medium risk (pH of 7.1–7.2 or
bicarbonate level of 6–10 mmol/L,
newly diagnosed T1D, or concern for
unreliable home care), or high risk
(pH , 7.1 or bicarbonate level # 5
mmol/L,16,17 Glasgow Coma Scale #
13 or abnormal neurologic
examination result,18,19 age #36
months,8,20 glucose level . 1000 mg/
dL,17 blood urea nitrogen level .
30 mg/dL,17,18,21 serum osmolality
level . 330 mOsm/kg,16 or
potassium level ,3 mEq/L16). We
excluded children who were
transferred to our institution from an
outside ED or hospital and children
who did not have DKA.
Interventions
Key Drivers
Through collaboration with key
stakeholders, we identified 3 key
drivers of care (Fig 1). To address our
key drivers, we implemented multiple
change strategies, including
development of an evidence-based
guideline (EBG), EMR order set
updates, regular updates to ED
physicians and stakeholders, and
nursing education.
Guideline Development and
Implementation
Our EBG (Supplemental Fig 4) was
developed by using
a multidisciplinary approach over the
course of 6 months by a guideline
implementation team that included
experts in the management of
children with T1D (pediatric
emergency medicine physicians,
pediatric endocrinologists, nursing
supervisors, ED nurses, and
pharmacists), a data and cost analysis
(health economist and statistician),
and a member of our hospital’s family
advisory council. Current medical
literature and available practice

FIGURE 1
Key driver diagram.
standards were reviewed extensively
and incorporated into our treatment
guideline. Our EBG was presented to
key stakeholders, including ED
physicians, endocrinologists,
intensive care physicians, and
hospitalists, on multiple occasions.
Feedback from stakeholders was used
to refine our treatment guideline
through an iterative process. Our EBG
was implemented in April 2016,
which included placement of the
guideline on our intranet clinical
guideline webpage, a nested link
within our ED DKA order set, and
printed version manually placed
within our 2 EDs to promote
accessibility.
EMR Order Set and Discharge Updates
Before our QI initiative, multiple
order sets specific to diabetes
evaluation were available to ED
providers. We reviewed all available
order sets and refined these to be
more specific to presenting concerns.
Through collaboration with
Information Technology, we created
an ED DKA order set, which is
consistent with our EBG. Key updates
included the following: embedding
our EBG within the order set for ease
of reference; integration of point-of-
care testing to shorten time to
laboratory results (ie, blood gas,
glucose, and electrolytes); options for
normal saline bolus amounts of
10 mL/kg if the patient was
#36 months or 20 mL/kg if the
patient was .36 months, given over
Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020
PEDIATRICS Volume 145, number 3, March 2020
1 hour; standard dosing of 0.1 U/kg
for administration of fast-acting
subcutaneous insulin; and automated
repeat point-of-care glucose checks
1 and 2 hours after insulin
administration. Given that we wanted
1 order set for all children with DKA,
including those with moderate or
severe (medium or high risk) DKA,
we updated our order set to include
treatment options for these patients.
In addition, we updated our ED
discharge instructions for sick-day
management for children treated with
subcutaneous insulin and
insulin pumps.
Nursing and Provider Education
Several nursing and provider change
strategies were used to promote
shared awareness of and adherence
to our EBG. We conducted 5 nursing
education sessions during June 2016
and July 2017 to refine nursing skills
and knowledge of T1D management,
specifically subcutaneous and IV
insulin administration, and IV fluid
management. These sessions included
case-based education using treatment
recommendations from our EBG,
hands-on training with trifuse IV
kits for fluid and medication
compatibility, education on insulin
drip management, and education on
signs of complications from insulin or
IV fluid therapy (ie, hypoglycemia,
altered mental status). To further
increase the success of our EBG
implementation, we held quarterly
presentations to ED physicians,
during which time interim results
were presented and key aspects of
the treatment guideline were
reinforced. Updates were provided to
endocrine physicians on an
annual basis.
Measures
Our primary outcome was the
proportion of children with low-risk
DKA who were hospitalized. Our
balancing measure was ED return
visits within 3 days of the index ED
visit. Secondary outcomes were the
proportion of children who received
rapid or fast-acting subcutaneous
insulin in the ED, ED length of stay
(LOS), and ED charges.
Study of the Intervention
The hospitalization rate was
monitored by using statistical process
control methods.22,23 Specifically, the
percentages of hospitalizations were
assessed via p-charts.
Analysis
Patient characteristics before
(January 1, 2012, to March 31, 2016)
and after implementation (April 1,
2016, to December 31, 2018) of the
QI initiative were compared.
Differences in numeric and
categorical data were assessed by
using the Mann–Whitney U test and
x2 test (or Fisher’s exact test),
respectively. Hemoglobin A1C was
reported as ,7.5% or $7.5%.24
Because of small sample sizes within
each quarter, sequential groups of 9
consecutive patients were used for
p-charts. Because of small sample
sizes within each sequential group,
the normal approximation to the
binomial distribution did not apply.
Therefore, control limits for p-charts
were calculated by using exact
binomial percentiles corresponding
to 2 and 3 SDs from the mean under
a normal distribution (ie, 2.3 and 97.7
percentiles and 0.1 and 99.9
percentiles, respectively). Trial limits
were calculated after 11 points,
corresponding to the
3
preimplementation period, and TABLE 1 Demographics for Children With Established T1D Presenting to the ED With Low-risk DKA
hospitalization rates were compared Preimplementation Postimplementation P
to this baseline. The centerline was January 1, 2012, to March 31, April 1, 2016 to December 31,
then shifted when special cause 2016 2018
variation was seen. Western Electric N = 99 N = 66
rules23 were used to indicate an out-
Age, y, median (IQR) 12 (9–15) 15 (10–17) .013
of-control process. A cumulative Male sex, n (%) 37 (37) 30 (46) .301
summation chart was used to Race and/or ethnicity, n (%) .124
estimate the number of avoided White 54 (55) 43 (65)
hospitalizations after implementation African American 34 (34) 14 (21)
Hispanic 3 (3) 2 (3)
with respect to the preintervention
American Indian 2 (2) 3 (5)
hospitalization rate.25 A post hoc . 1 race 2 (2) 4 (6)
logistic regression was performed to Other, unknown, or 4 (4) 0 (0)
account for potential confounders declined
that might impact hospitalization, Insurance status, n (%) .602
Commercial 54 (55) 36 (55)
including age, sex, race, insurance
Medical assistance 33 (33) 23 (35)
status, hemoglobin A1C at ED Medicare or Medicaid 12 (12) 6 (9)
presentation, and implementation Self-pay 0 (0) 1 (1)
period. Total ED and hospital charges Hemoglobin A1C, n (%)a .360
were reported as medians with ,7.5 6 (6) 2 (3)
$7.5 88 (94) 62 (97)
interquartile ranges (IQRs) and were
a Seven encounters had missing values for hemoglobin A1C.
adjusted to 2018 dollars by using the
Bureau of Labor Statistics Consumer
Price Index for hospital and related postimplementation period (Table 2). unadjusted analysis
services. The total adjusted ED Children presenting in the (preimplementation: 75%[ 95% CI
charges represent all ED charges, postimplementation period were 65%–83%]; postimplementation:
excluding any hospital facility charges significantly older than children in 54% [95% CI 42%–67%])
from an ED visit. The total adjusted the preimplementation period (P = (Supplemental Table 4). The
charges represent all facility and .013). An adjusted post hoc logistic hospitalization rate in the
professional charges for a given regression analysis revealed that postimplementation period revealed
encounter. All analyses were hospitalization rates were special cause variation between
performed in SAS (SAS Institute, Inc, comparable with those in the sequential groups 3 and 4 (Fig 2).
Cary, NC) or R (R Core Team, R
Foundation for Statistical Computing, TABLE 2 Pre- and Postimplementation Outcomes and Resource Use for Children With Established
Vienna, Austria). T1D Presenting to the ED With Low-risk DKA
Preimplementation Postimplementation P
January 1, 2012, to March April 1, 2016, to December
RESULTS 31, 2016 31, 2018
We identified 3132 total T1D-related N = 99 N = 66
ED encounters, of which 974 involved Hospital admission, n (%) 73 (74) 36 (55) .011
children with DKA. We excluded 381 ED return visits, n (%)
and 236 encounters for moderate and Within 3 d 0 (0) 1 (2) .400
severe DKA, respectively. We further Within 7 d 2 (2) 2 (3) .219
Within 14 d 5 (5) 2 (3) .528
excluded 136 encounters in which the
Subcutaneous rapid or fast-acting 33 (33) 23 (35) .840
child presented with newly diagnosed insulin given in ED, n (%)
T1D and 56 encounters resulting Median NS bolus, mL/kg, (IQR) 10 (10–10) 12.5 (10–19) ,.001
from transfers from another hospital. Median ED LOS, min, (IQR) 215 (180–263) 241 (195–300) .031
This resulted in a study cohort of 165 Median hospital LOS, h, (IQR) 23 (0–36) 18 (0–28) .066
Total adjusted median ED charges, 1680 (1485–1833) 1546 (1387–1701) .002
children with low-risk DKA (Table 1).
$a, (IQR)
Total adjusted median hospital 11 942 (4144–16 518) 10 374 (3072–14 238) .121
The baseline hospitalization rate in charges, $a, (IQR)
the preimplementation period was
74% (95% confidence interval [CI] NS, normal saline.
a Adjusted for inflation by using the Consumer Price Index for hospital and related services for 2018 dollars. Total
64%–82%), and this decreased to adjusted ED charges represent all ED charges, excluding any hospital facility charges from an ED visit. Total adjusted
55% (95% CI 42%–67%) in the charges represent all facility and professional charges for a given encounter.

Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020

4 BERGMANN et al
 compared with $3102 (SD $2889) for
those discharged from the ED (mean
difference: $13 022; 95% CI
$11 418–$14 627). The most common
reason for hospitalization was
initiation of IV insulin (Table 3).

DISCUSSION
Through implementation of a QI
initiative, we safely reduced
hospitalization rates by an absolute
difference of 19% for children with
low-risk DKA, without an increase in
3-day ED revisits. Special cause
variation was seen after
implementation of a nursing
algorithm and nursing education,
suggesting that these efforts
influenced practice change. The total
FIGURE 2 adjusted mean charges were $13 022
Statistical process control chart revealing hospitalization rates by sequential groups of 9 consec- higher among children hospitalized
utive children presenting to the ED with low-risk DKA. Yellow and red control limits (CLs) correspond compared with those discharged from
to exact binomial percentiles corresponding to 2 and 3 SDs from the mean under a normal
distribution. the ED.

Children with DKA are often

After special cause variation was Table 5) compared with those hospitalized,5 leading to substantial
seen, there was a 37–percentage discharged from the ED. The total medical expenditure. Recent
point (95% CI 14%–59%) decrease in adjusted mean charge was $16 125 investigation of insurance claims data
hospitalizations (P , .001). A (SD $14 534) for those hospitalized, suggests that inpatient care accounts
cumulative summation analysis
revealed that hospitalizations
began to decrease after EBG
implementation, and this trend
continued throughout the QI initiative
(Fig 3). Thirteen potential
hospitalizations were avoided, which
surpassed our goal of a 10%
reduction (ie, a 10% reduction
corresponds 7 avoided
hospitalizations). The median ED LOS
was significantly longer in the
postimplementation period
(241 minutes; IQR 195–300
minutes) compared with the
preimplementation period (215
minutes; IQR 180–263 minutes; P =
.031) (Table 2). The number of
children with 3-day ED revisits did
not change significantly.

In the postimplementation period,

children who were hospitalized were FIGURE 3
similar in median age, pH, and Cumulative number of hospitalizations relative to the preimplementation period for children with
bicarbonate levels (Supplemental established T1D presenting to the ED with low-risk DKA. Q, quarter.

Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020

PEDIATRICS Volume 145, number 3, March 2020 5
TABLE 3 Reasons for Hospital Admission Among Children Presenting With Low-risk DKA in the
Postimplementation Period
Category
Provider discretion
Unreliable home care
Comorbid illness
a Persistent vomiting despite antiemetics and IV fluids.
for .40% of total annual medical
expenditures for youth with DKA,
which is estimated to be $14 200 for
those with 1 episode of DKA
compared with $8400 for those with
no episodes, an excess of $5800.6 This
is similar to results from a large study
across 38 US children’s hospitals in
which the total cost of hospitalization
for DKA was estimated to be $7160
per encounter.5 Our results suggest
that reducing admissions for children
with low-risk DKA may lead to
substantial cost savings.
As the number of children with T1D
rises, it is increasingly important to
develop management strategies for
diabetes-related complications.
Previous investigation has revealed
that children with mild DKA have
resolution of acidosis within 3 to
7 hours,11,14,26 suggesting that
hospitalization may be avoidable in
this population. One recent pre- and
postimplementation study revealed
that hospitalizations and 30-day
readmissions for children with DKA
decreased after standardization of
T1D education, increasing intensive
insulin regimens, and community
engagement.27 However, this study
was conducted in a diabetes
ambulatory clinic, limiting
generalizability to an ED
population.27 Additional efforts in
ambulatory settings have been
focused on telephone follow-up,28–30
text-messaging support,31 and use of
Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020
6 BERGMANN et al
Reason for Hospitalization (N = 36)
IV insulin initiated (n = 17)
Per endocrine recommendation (n = 4 of 17)
ED provider recommended hospitalization (n = 4)
Poor insulin knowledge, unreliable glucose checks (n = 4)
Multiple previous DKA episodes (n = 1)
Poor control per recent A1C values (n = 1)
Social concerns (n = 1)
Dehydration (n = 4)a
Acute kidney injury (n = 1)
Asthma exacerbation (n = 1)
Seizure (n = 1)
Intractable migraine (n = 1)
a diabetes hotline.32 However, the
impact on hospitalization with these
interventions is unclear. One
retrospective study revealed that use
of a diabetes hotline was associated
with decreased frequency of DKA,32
whereas a more rigorous study in
which patients were randomly
assigned to scheduled telephone
support compared with usual care
revealed no difference in
hospitalization rates for DKA.29
Current treatment guidelines note
that children with established T1D
and reliable home care who present
with hyperglycemia and ketosis
without vomiting or severe
dehydration may be managed at
home or an outpatient setting.8
However, there are no specific
recommendations for children who
present with mild DKA for whom
discharge may be appropriate. As
a result, there remains uncertainty
about which treatment options are
optimal for this population. For
instance, treatment with high fluid
volumes (20 mL/kg of isotonic bolus
1 1.5 3 maintenance rate) has been
shown to shorten time to metabolic
normalization compared with
treatment with lower fluid volumes
(10 mL/kg of isotonic bolus 1 1.25 3
maintenance rate)33 and may be more
appropriate for children with mild
DKA for whom discharge from the ED
is anticipated. Importantly, the
incidence of clinically apparent brain
injury and cognitive function after
recovery from DKA have been shown
to be similar among children
randomly assigned to receive larger
(20 mL/kg) compared with smaller
volumes of isotonic fluid boluses
(10 mL/kg).34 During our QI
initiative, the median isotonic fluid
bolus volume increased significantly
in the postimplementation period,
suggesting adherence to our EBG and
adoption of this practice among ED
providers. In addition, providers who
care for children with DKA must
consider risk factors that may
influence provision of adequate home
care. This includes parent and child
knowledge of sick-day management,
psychological comorbidities,
availability of home insulin, and
blunting of further dehydration.
Implementation of an EBG like the
one we developed leads to
standardization of care and reduced
treatment variation and may guide
providers who treat such children
with less frequency.
We encountered many challenges
when implementing our QI initiative.
First, some providers were concerned
about discharging patients who were
mildly acidotic and might have
progression of illness at home and
therefore return with more severe
DKA. This led to our classification of
low-risk DKA, which includes
a combination of laboratory
parameters and clinical variables that
are used to identify patients who are
unlikely to have progression of
illness. Second, our guideline was
developed with the intent of
prolonging isotonic IV fluid
administration for low-risk patients,
which led to concern that our EBG
would increase ED LOS and constrain
bed space. Although ED LOS did
increase in the postimplementation
period, this was easy for providers to
accept because low-risk patients with
DKA present relatively infrequently.
Third, we found it necessary to
regularly increase provider
awareness of our EBG. We held
regular educational presentations and
updates at staff meetings and sent
electronic reminders to providers. For
the few providers who were hesitant
to use our EBG in practice, this
provided an opportunity to review
cases, which helped to obtain buy-in.
Fourth, we found that administration
of subcutaneous insulin in the ED did
not change over the QI initiative. This
may have been due to treatment with
home insulin shortly before ED
presentation, treatment with home
insulin pens in the ED, deferring
insulin treatment to the admitting
provider, or provider discretion. This
represents an area on which we can
further improve care.
Our study has several limitations.
First, we could not evaluate for
secular trends because our study
population was defined by laboratory
and clinical criteria that are not
included in large data sets from US
children’s hospitals. Second, it is
possible that children may have
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Funded in part by the Children’s Hospital Association.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
REFERENCES
1. Mayer-Davis EJ, Dabelea D, Lawrence
JM. Incidence trends of type 1 and type
2 diabetes among youths, 2002-2012.
N Engl J Med. 2017;377(3):301
2. Dabelea D, Mayer-Davis EJ, Saydah S,
et al; SEARCH for Diabetes in Youth
Study. Prevalence of type 1 and type 2
diabetes among children and
adolescents from 2001 to 2009. JAMA.
2014;311(17):1778–1786
3. Dabelea D, Rewers A, Stafford JM, et al;
SEARCH for Diabetes in Youth Study
Group. Trends in the prevalence of
ketoacidosis at diabetes diagnosis: the
SEARCH for diabetes in youth study.
Pediatrics. 2014;133(4). Available at:
www.pediatrics.org/cgi/content/full/
133/4/e938
4. Rewers A, Klingensmith G, Davis C, et al.
Presence of diabetic ketoacidosis at
Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020
PEDIATRICS Volume 145, number 3, March 2020
returned to another ED, leading to an
underestimate of ED revisits.
However, this is unlikely because
children with T1D are managed
closely by our endocrinology clinic
and often have next-day follow-up.
Third, although we found a 19%
absolute reduction in the
hospitalization rate, special cause
variation was not seen until
sequential groups 3 to 4 after
implementation of our initiative,
around the time of the second ED
nursing education session. Because
of this, we introduced a process
change after sequential group 3.
Although this analysis revealed
a 37–percentage point reduction in
hospitalizations, we feel that further
data collection is needed to
determine if this improvement is
sustained. Finally, our results may
not be generalizable to hospitals
without a robust T1D program.
Future efforts may be focused on
provider-driven variations in care and
diagnosis of diabetes mellitus in youth:
the Search for Diabetes in Youth Study.
Pediatrics. 2008;121(5). Available at:
www.pediatrics.org/cgi/content/full/
121/5/e1258
5. Tieder JS, McLeod L, Keren R, et al;
Pediatric Research in Inpatient Settings
Network. Variation in resource use and
readmission for diabetic ketoacidosis
in children’s hospitals. Pediatrics. 2013;
132(2):229–236
6. Shrestha SS, Zhang P, Barker L,
Imperatore G. Medical expenditures
associated with diabetes acute
complications in privately insured U.S.
youth. Diabetes Care. 2010;33(12):
2617–2622
7. Della Manna T, Steinmetz L, Campos PR,
et al. Subcutaneous use of a fast-acting
insulin analog: an alternative treatment
implementation of telephone follow-
up with our EBG.
CONCLUSIONS
Through implementation of a QI
initiative, we safely reduced
hospitalizations for children with
established T1D presenting to our ED
with low-risk DKA. Implementation of
a QI initiative such as ours may lead
to substantial cost savings.
ABBREVIATIONS
CI: confidence interval
DKA: diabetic ketoacidosis
EBG: evidence-based guideline
ED: emergency department
EMR: electronic medical record
IQR: interquartile range
IV: intravenous
LOS: length of stay
QI: quality improvement
T1D: type 1 diabetes
for pediatric patients with diabetic
ketoacidosis. Diabetes Care. 2005;28(8):
1856–1861
8. Wolfsdorf JI, Glaser N, Agus M, et al.
ISPAD Clinical Practice Consensus
Guidelines 2018: diabetic ketoacidosis
and the hyperglycemic hyperosmolar
state. Pediatr Diabetes. 2018;19(suppl
27):155–177
9. Wolfsdorf J, Glaser N, Sperling MA;
American Diabetes Association. Diabetic
ketoacidosis in infants, children, and
adolescents: a consensus statement
from the American Diabetes
Association. Diabetes Care. 2006;29(5):
1150–1159
10. Bonadio WA. Pediatric diabetic
ketoacidosis: pathophysiology and
potential for outpatient management of
7
 selected children. Pediatr Emerg Care.
1992;8(5):287–290
11. Bonadio WA, Gutzeit MF, Losek JD, Smith
DS. Outpatient management of diabetic
ketoacidosis. Am J Dis Child. 1988;
142(4):448–450
12. Umpierrez GE, Cuervo R, Karabell A, Latif
K, Freire AX, Kitabchi AE. Treatment of
diabetic ketoacidosis with
subcutaneous insulin aspart. Diabetes
Care. 2004;27(8):1873–1878
13. Umpierrez GE, Latif K, Stoever J, et al.
Efficacy of subcutaneous insulin lispro
versus continuous intravenous regular
insulin for the treatment of patients
with diabetic ketoacidosis. Am J Med.
2004;117(5):291–296
14. Cohen M, Leibovitz N, Shilo S,
Zuckerman-Levin N, Shavit I, Shehadeh
N. Subcutaneous regular insulin for the
treatment of diabetic ketoacidosis in
children. Pediatr Diabetes. 2017;18(4):
290–296
15. Vincent M, Nobécourt E. Treatment of
diabetic ketoacidosis with
subcutaneous insulin lispro: a review of
the current evidence from clinical
studies. Diabetes Metab. 2013;39(4):
299–305
16. Edge JA, Jakes RW, Roy Y, et al. The UK
case-control study of cerebral oedema
complicating diabetic ketoacidosis in
children. Diabetologia. 2006;49(9):
2002–2009
17. Lawrence SE, Cummings EA, Gaboury I,
Daneman D. Population-based study of
incidence and risk factors for cerebral
edema in pediatric diabetic
ketoacidosis. J Pediatr. 2005;146(5):
688–692
18. Glaser NS, Marcin JP, Wootton-Gorges
SL, et al. Correlation of clinical and
biochemical findings with diabetic
ketoacidosis-related cerebral edema in
children using magnetic resonance
Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020
8 BERGMANN et al
diffusion-weighted imaging. J Pediatr.
2008;153(4):541–546
19. Marcin JP, Glaser N, Barnett P, et al;
American Academy of Pediatrics
Pediatric Emergency Medicine
Collaborative Research Committee.
Factors associated with adverse
outcomes in children with diabetic
ketoacidosis-related cerebral edema.
J Pediatr. 2002;141(6):793–797
20. Edge JA, Ford-Adams ME, Dunger DB.
Causes of death in children with insulin
dependent diabetes 1990-96. Arch Dis
Child. 1999;81(4):318–323
21. Glaser N, Barnett P, McCaslin I, et al;
Pediatric Emergency Medicine
Collaborative Research Committee of
the American Academy of Pediatrics.
Risk factors for cerebral edema in
children with diabetic ketoacidosis. The
Pediatric Emergency Medicine
Collaborative Research Committee of
the American Academy of Pediatrics.
N Engl J Med. 2001;344(4):264–269
22. Benneyan JC, Lloyd RC, Plsek PE.
Statistical process control as a tool for
research and healthcare improvement.
Qual Saf Health Care. 2003;12(6):
458–464
23. Montgomery DC. Introduction to
Statistical Quality Control, 7th ed.
Hoboken, NJ: John Wiley & Sons; 2009
24. American Diabetes Association. 12.
Children and adolescents: Standards of
Medical Care in Diabetes-2018. Diabetes
Care. 2018;41(suppl 1):S126–S136
25. Grigg OA, Farewell VT, Spiegelhalter DJ.
Use of risk-adjusted CUSUM and RSPRT
charts for monitoring in medical
contexts. Stat Methods Med Res. 2003;
12(2):147–170
26. Linares MY, Schunk JE, Lindsay R.
Laboratory presentation in diabetic
ketoacidosis and duration of therapy.
Pediatr Emerg Care. 1996;12(5):347–351
27. Ilkowitz JT, Choi S, Rinke ML, Vandervoot
K, Heptulla RA. Pediatric type 1 diabetes:
reducing admission rates for diabetes
ketoacidosis. Qual Manag Health Care.
2016;25(4):231–237
28. Lawson ML, Cohen N, Richardson C,
Orrbine E, Pham B. A randomized trial
of regular standardized telephone
contact by a diabetes nurse educator in
adolescents with poor diabetes control.
Pediatr Diabetes. 2005;6(1):32–40
29. Nunn E, King B, Smart C, Anderson D. A
randomized controlled trial of
telephone calls to young patients with
poorly controlled type 1 diabetes.
Pediatr Diabetes. 2006;7(5):254–259
30. Howells L, Wilson AC, Skinner TC,
Newton R, Morris AD, Greene SA. A
randomized control trial of the effect of
negotiated telephone support on
glycaemic control in young people with
type 1 diabetes. Diabet Med. 2002;19(8):
643–648
31. Franklin VL, Waller A, Pagliari C, Greene
SA. A randomized controlled trial of
Sweet Talk, a text-messaging system to
support young people with diabetes.
Diabet Med. 2006;23(12):1332–1338
32. Chiari G, Ghidini B, Vanelli M.
Effectiveness of a toll-free telephone
hotline for children and adolescents
with type 1 diabetes. A 5-year study.
Acta Biomed. 2003;74(suppl 1):45–48
33. Bakes K, Haukoos JS, Deakyne SJ, et al.
Effect of volume of fluid resuscitation
on metabolic normalization in children
presenting in diabetic ketoacidosis:
a randomized controlled trial. J Emerg
Med. 2016;50(4):551–559
34. Kuppermann N, Ghetti S, Schunk JE,
et al; PECARN DKA FLUID Study Group.
Clinical trial of fluid infusion rates for
pediatric diabetic ketoacidosis. N Engl
J Med. 2018;378(24):2275–2287
 A Quality Improvement Initiative to Reduce Hospitalizations for Low-risk
Diabetic Ketoacidosis
Kelly R. Bergmann, M. Jennifer Abuzzahab, Joe Arms, Gretchen Cutler, Heidi
Vander Velden, Ted Simper, Eric Christensen, Dave Watson and Anupam Kharbanda
Pediatrics originally published online February 13, 2020;

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/early/2020/02/11/peds.2
019-1104
References This article cites 31 articles, 7 of which you can access for free at:
http://pediatrics.aappublications.org/content/early/2020/02/11/peds.2
019-1104#BIBL
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Emergency Medicine
http://www.aappublications.org/cgi/collection/emergency_medicine_
sub
Endocrinology
http://www.aappublications.org/cgi/collection/endocrinology_sub
Diabetes Mellitus
http://www.aappublications.org/cgi/collection/diabetes_mellitus_sub
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or
in its entirety can be found online at:
http://www.aappublications.org/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020

 A Quality Improvement Initiative to Reduce Hospitalizations for Low-risk
Diabetic Ketoacidosis
Kelly R. Bergmann, M. Jennifer Abuzzahab, Joe Arms, Gretchen Cutler, Heidi
Vander Velden, Ted Simper, Eric Christensen, Dave Watson and Anupam Kharbanda
Pediatrics originally published online February 13, 2020;

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://pediatrics.aappublications.org/content/early/2020/02/11/peds.2019-1104

Data Supplement at:

http://pediatrics.aappublications.org/content/suppl/2020/02/12/peds.2019-1104.DCSupplemental

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since 1948. Pediatrics is owned, published, and trademarked by
the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,
60007. Copyright © 2020 by the American Academy of Pediatrics. All rights reserved. Print
ISSN: 1073-0397.

Downloaded from www.aappublications.org/news at Bibliotheque Univ Paris V on February 20, 2020