BACKGROUND:

Given the risks associated with antibiotics, efforts to reduce unnecessary antibiotic use in the NICU have become increasingly urgent. In 2016, a comprehensive 3-year quality improvement (QI) initiative was conducted in a level 4 NICU that sought to decrease the antibiotic use rate (AUR) by 20%.

METHODS:

This local QI initiative was conducted in the context of a multicenter learning collaborative focused on decreasing unnecessary antibiotic use. Improvement strategies focused on addressing gaps in the core elements of antibiotic stewardship programs. Outcome measures included the AUR and the percent of infants discharged without antibiotic exposure. Process measures included the percent of infants evaluated for early-onset sepsis (EOS) and duration of antibiotics used for various infections. Statistical process control charts were used to display and analyze data over time.

RESULTS:

The AUR decreased from 27.6% at baseline to 15.5%, a 43% reduction, and has been sustained for >18 months. Changes most attributable to this decrease include implementation of the sepsis risk calculator, adopting a 36-hour rule-out period for sepsis evaluations, a 36-hour antibiotic hard stop, and novel guideline for EOS evaluation among infants <35 weeks. The percent of infants discharged without antibiotic exposure increased from 15.8% to 35.1%. The percent of infants ≥36 weeks undergoing evaluation for EOS decreased by 42.3% and for those <35 weeks by 26%.

CONCLUSIONS:

Our efforts significantly reduced antibiotic use and exposure in our NICU. Our comprehensive, rigorous approach to QI is applicable to teams focused on improvement.

Antibiotics are powerful, life-saving drugs, but when used inappropriately, they may have serious adverse effects. For example, prolonged empirical antibiotic use among preterm infants with negative culture results has been associated with an increased risk of mortality and major morbidities, including late-onset sepsis, necrotizing enterocolitis (NEC), and stage 3 or greater retinopathy of prematurity.14  Additional consequences of antibiotic use include emergence of fungal infections and multidrug-resistant bacteria.5,6  More recently, the influence of antibiotics on the developing microbiome, and potentially important health outcomes, has been an important focus, particularly among neonates.7  Despite these concerns, antibiotic use varies widely among NICUs, particularly in the absence of proven infection, and as a result, some antibiotic use is unnecessary.8,9  Thus, antibiotic stewardship is necessary to reduce overuse and avoid potential adverse outcomes.

Whereas our global aim was to optimize antibiotic use as part of a comprehensive antibiotic stewardship program, our specific aim for this initiative was to decrease unnecessary antibiotic use through involvement in the Vermont Oxford Network’s Internet-based Newborn Improvement Collaborative for Quality (iNICQ) entitled “Choosing Antibiotics Wisely,” which was conducted from January 2016 through December 2018. Given published evidence suggesting some antibiotic use in the NICU is unnecessary,8,9  alongside knowledge of local practices that were likely contributing to overuse (eg, prolonged treatment of well-appearing newborns based on risk factors and/or laboratory values despite negative culture results), we chose to focus on reducing antibiotic use. At the outset, we sought to decrease our antibiotic use rate (AUR), defined as the percent of days patients received ≥1 antibiotic, from 27.6% to 22.0% (a 20% reduction) by December 2016. Once this was achieved, we believed additional improvement was possible and updated our aim statement to reflect our goal of achieving an additional 10% decrease in the AUR (from 18.9% to 17.0%) by December 2018.

The Golisano Children’s Hospital NICU is a 68-bed level 4 regional perinatal center in Rochester, New York, with a single-patient room design and nearly 1100 admissions yearly, 170 of whom are very low birth weight. Our unit provides the highest level of care for critically ill newborns in the Finger Lakes region of New York state, including those with congenital heart disease and those requiring advanced surgical services, therapeutic hypothermia, and extracorporeal membrane oxygenation. At baseline, ∼5% to 10% of all admissions to the NICU were solely for an early-onset sepsis (EOS) evaluation, including asymptomatic newborns born to mothers with chorioamnionitis, because all sepsis evaluations in newborns occurred in the NICU.

To begin, an interprofessional team was gathered that included neonatologists, neonatal fellows, advanced practice providers (APPs), nurses from the NICU and well-infant nursery, newborn hospitalists, pharmacists, information analysts, pediatric infectious diseases specialists, nurse educators, and 2 parents. Additional representation was sought when relevant to the work being done (eg, pediatric surgery involvement for discussing antibiotic use in infants born with gastroschisis). This team participated in periodic webinars as part of the iNICQ learning collaborative and met monthly to review data, develop and discuss change ideas, and foster collaboration.

The team reviewed baseline data and created a key driver diagram to develop a shared mental model for our approach to reducing unnecessary antibiotic use (Fig 1). Smaller work groups formed, typically consisting of a neonatologist, fellow or APP, infectious disease specialist, and nurse, to facilitate greater participation by all team members. Strategies for improvement included frequent audits and feedback, use of plan-do-study-act cycles, and standardization of practices through development of several written guidelines (Table 1).

FIGURE 1

Key driver diagram of reducing unnecessary antibiotic use linking change ideas to primary and secondary drivers. CRP, C-reactive protein; ID, infectious diseases.

FIGURE 1

Key driver diagram of reducing unnecessary antibiotic use linking change ideas to primary and secondary drivers. CRP, C-reactive protein; ID, infectious diseases.

TABLE 1

Timeline of Interventions

DateDescription of Test of Change and/or Intervention
April 2016 Post-AUR in NICU 
April to June 2016 SRC pilot and testing on 10 infants, education of obstetrics nursing staff 
July 2016 SRC used for all infants ≥36 wk’ gestation 
August 2016 Adoption of 36-h rule-out period for all sepsis evaluations 
November 2016 36-h hard stop in EMR for all antibiotic orders 
January 2017 Provider-specific AUR report created and given to neonatologists 
February 2017 Guideline for congenital pneumonia 
March 2017 Guideline for catheter-associated phlebitis 
May 2017 Guideline for UTI 
June 2017 SRC used for all infants ≥35 wk’ gestation 
August 2017 Guideline for antibiotic use in infants with gastroschisis; EOS guideline for infants born <35 wk’ gestation 
March 2018 Scrub the hub of all central lines with alcohol 
June 2018 Use of blue microclaves on all central line access points; guideline for evaluation and management of NEC 
DateDescription of Test of Change and/or Intervention
April 2016 Post-AUR in NICU 
April to June 2016 SRC pilot and testing on 10 infants, education of obstetrics nursing staff 
July 2016 SRC used for all infants ≥36 wk’ gestation 
August 2016 Adoption of 36-h rule-out period for all sepsis evaluations 
November 2016 36-h hard stop in EMR for all antibiotic orders 
January 2017 Provider-specific AUR report created and given to neonatologists 
February 2017 Guideline for congenital pneumonia 
March 2017 Guideline for catheter-associated phlebitis 
May 2017 Guideline for UTI 
June 2017 SRC used for all infants ≥35 wk’ gestation 
August 2017 Guideline for antibiotic use in infants with gastroschisis; EOS guideline for infants born <35 wk’ gestation 
March 2018 Scrub the hub of all central lines with alcohol 
June 2018 Use of blue microclaves on all central line access points; guideline for evaluation and management of NEC 

The AUR, a new measure for our unit, was calculated both weekly and monthly and posted in several NICU locations to provide a visual display of improvement over time. Because many infants were being admitted solely for sepsis evaluation, we pursued a staged implementation of the Kaiser sepsis risk calculator (SRC).10  Several tests of change were conducted over a 3-month period that included retrospective calculations for term infants born to mothers with chorioamnionitis (92% of whom had a sepsis risk score <0.7), piloting use of the SRC in 10 term infants born to mothers with chorioamnionitis to determine if antibiotics should be initiated, feedback to NICU staff on appropriate use and interpretation of the SRC, and education of obstetrics nursing staff on use of the SRC, which included sample cases for practice. In July 2016, we began using the SRC for all infants born ≥36 weeks to determine, with the clinical examination, if an evaluation for sepsis was warranted. We chose 36 weeks to mirror a similar approach shared from another iNICQ team who safely implemented use of the SRC in this population. After a year of experience, we expanded the SRC for all infants born ≥35 weeks.

Published and local microbiology data suggested that stopping antibiotics at 36 hours when blood culture results were negative would be safe and result in fewer administered doses of antibiotics, so this practice was adopted in August 2016.11,12  To ensure reliability of a sepsis evaluation, staff were reeducated on desired blood volume for culture (≥1.5 mL). A mandatory field was also created to document blood volume for culture so providers would be confident in negative results when an adequate volume was obtained. Review of local data also suggested that as many as 10% to 15% of all infants being treated for possible sepsis in our NICU received additional doses of antibiotics because orders were not discontinued in a timely manner. As a result, we implemented a 36-hour hard stop for all antibiotics in the electronic medical record (EMR) in November 2016.

After reviewing the results of an audit conducted through the iNICQ learning collaborative, we identified several gaps in practice in which improvements were possible to decrease unnecessary antibiotic use. These included developing written guidelines for common infections, providing provider-specific feedback on prescribing patterns, and reducing antibiotic use in preterm infants. Early in 2017, we developed a report displaying provider-specific AURs that was circulated to every neonatologist for calendar year 2016. The report was anonymous, with the exception that each physician was made aware of their own AUR, and demonstrated the variability in AUR among neonatologists (Supplemental Fig 6). The report was updated at 6-month intervals to provide ongoing feedback because we believed benchmarking against colleagues would motivate a change in prescribing practices. A detailed review of local practices suggested that variation was occurring in the management of several common infections. Duration of therapy ranged from 7 to 10 days in infants with congenital pneumonia. Diagnostic criteria used for urinary tract infection (UTI), including colony counts, results of urinalysis, and method of specimen collection, also varied widely. A chart review of infants with catheter-associated phlebitis revealed that 100% of these infants were exposed to at least 2 days of antibiotics, although none had bacteremia or were clinically ill. Lastly, wide variation in antibiotic use was observed in infants with gastroschisis after abdominal wall closure.

In February 2017, a guideline for congenital pneumonia was adopted that included an algorithm to guide decision-making on cessation of antibiotics at 36 hours, 5 days, or at most 7 days on the basis of resolution of symptoms. The following month, our unit adopted a new guideline for the management of catheter-associated phlebitis including the use of warm compresses without antibiotics while maintaining the catheter. In May 2017, a guideline for diagnosis and management of UTI was implemented on the basis of input from neonatology, pediatric infectious diseases, and pediatric nephrology. This guideline emphasizes obtaining a reliable culture and provides guidance for colony counts that should be considered significant and pathogenic. After collaboration with pediatric surgery, a new guideline was developed in August 2017 for infants born with gastroschisis that emphasized cessation of antibiotics within 24 hours of abdominal wall closure, by using the unit’s approach to EOS in the decision to begin antibiotics after birth, and recognized that antibiotics were not indicated solely on the basis of the presence of bowel in a silo.

At baseline, >90% of infants born <35 0/7 weeks’ gestation received antibiotics in our NICU as part of an evaluation for EOS. However, published data suggest that preterm infants are at low risk for infection if they are delivered without risk factors (ie, delivery for maternal indications without chorioamnionitis, labor, or rupture of membranes).13,14  Moreover, local data suggested that 15% to 20% of infants born <35 0/7 weeks were delivered without risk factors for infection; none had bacteremia over a 5-year period, yet 85% received empirical antibiotics. As a result, a guideline for evaluation and management of EOS was developed in August 2017 for infants born <35 0/7 weeks that stratifies infants as either low risk or higher risk on the basis of risk factors mentioned above. Infants deemed low risk do not receive empirical antibiotics, and stopping empirical antibiotics is encouraged after 36 hours if culture results remain negative in those who are at higher risk.

An increase in observed central line–associated bloodstream infections (CLABSIs) in late 2017, associated with a concomitant increase in the AUR, prompted a focus on infection prevention in 2018. Two new practices were trialed and subsequently adopted to reduce risk of infection: scrubbing the hub of all central lines with alcohol and the use of blue microclaves on all central line access points. A new guideline was developed for evaluation and management of NEC that provided guidance for antibiotic choice and duration on the basis of the modified Bell’s stage to further standardize practice.

The primary outcome measure was the AUR, represented as the number of days patients received ≥1 antibiotic per 100 patient-days and displayed as a percentage on time-ordered charts. Using this definition of AUR reduced the burden of more complex data gathering and is consistent with that used by the California Perinatal Quality Care Collaborative.15 Our secondary outcome measure was the percentage of infants discharged without receiving antibiotics while in the NICU. Process measures included the percentage of infants born ≥36 weeks’ gestation undergoing a sepsis evaluation in the first 72 hours of life (total infants ≥36 weeks with a blood culture sent divided by total live births ≥36 weeks), percentage of infants <35 weeks’ gestation receiving empirical antibiotics within the first 72 hours of life (total infants <35 weeks with a blood culture sent divided by total live births <35 weeks), and duration of antibiotic use for infections that reflected compliance with new guidelines. Our balancing measures were cases of missed bacteremia, ascertained by reviewing all blood cultures obtained during the first week of life and determining if any delay in diagnosis and/or treatment resulted from the new approach to EOS, nosocomial sepsis rate, and infection-related mortality. All data were gathered from a local database and/or extracted from the EMR. Statistical process control charts (QI Macros, KnowWare International, Inc, Denver, CO) were used to display and analyze data over time. Special cause variation was determined by using rules suggested for health care.16 

This project was considered exempt by the institutional review board because it represented quality improvement (QI).

After staged implementation of the SRC, a significant decline in the percentage of infants ≥36 weeks undergoing a sepsis evaluation in the first 72 hours of life was observed (Fig 2), which has been sustained over time. Coinciding with this decline in sepsis evaluations was a concomitant decline in the AUR (Fig 3). Special cause variation was first noted beginning in July 2016, with 8 or more consecutive points below the baseline mean, resulting in a decrease in the AUR by 31.5% to 18.9%. This shift was attributed to implementation of the SRC, adoption of a 36-hour rule-out period for all sepsis evaluations, and use of a 36-hour hard stop in the EMR for all antibiotic orders. Introduction of new guidelines for common infections early in 2017 likely did not have a significant impact on the AUR because only common cause variation was noted through most of 2017. After implementation of a guideline for EOS in infants <35 weeks, we anticipated observing a further decline in the AUR. The average percentage of infants <35 weeks receiving empirical antibiotics for EOS declined after implementation of the SRC from 94.9% to 84.9%, suggesting that focusing on reducing unnecessary antibiotics for EOS in a separate patient population impacted antibiotic use in this population (Fig 4). After implementation of the guideline for EOS in infants <35 weeks, the average percent of infants being evaluated for EOS each month decreased further from 84.9% to 70.2%, a 17% relative decrease and a total relative decrease of 26% from the initial baseline. However, this did not result in an immediate decline in the observed AUR. As noted, a spike in the AUR, representing special cause variation, was observed in the AUR in October and November 2017, which coincided with an increase in CLABSIs (5 over a 3-week period) in the setting of an average daily census of 80 in our 68-bed unit for the entire month of October. Once infections were treated, our AUR declined below our baseline average, and special cause variation was again noted. The AUR declined by an additional 18% to a monthly average of 15.5% and has been sustained for >18 months.

FIGURE 2

P-chart displaying the percent of infants born ≥36 0/7 weeks being evaluated for EOS (total live births = 11 736). Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL). CL, center line (mean).

FIGURE 2

P-chart displaying the percent of infants born ≥36 0/7 weeks being evaluated for EOS (total live births = 11 736). Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL). CL, center line (mean).

FIGURE 3

U-prime chart displaying monthly AUR (total patient-days =120 113). Annotations indicate where tests of change occurred relative to the AUR. The center line (CL) (mean) was adjusted after special cause variation occurred on 2 separate occasions. Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL).

FIGURE 3

U-prime chart displaying monthly AUR (total patient-days =120 113). Annotations indicate where tests of change occurred relative to the AUR. The center line (CL) (mean) was adjusted after special cause variation occurred on 2 separate occasions. Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL).

FIGURE 4

P-chart displaying the percent of infants born <35 0/7 weeks receiving empirical antibiotics after birth (total live births = 1741). The center line (CL) (mean) was adjusted after special cause variation occurred on 2 separate occasions. Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL).

FIGURE 4

P-chart displaying the percent of infants born <35 0/7 weeks receiving empirical antibiotics after birth (total live births = 1741). The center line (CL) (mean) was adjusted after special cause variation occurred on 2 separate occasions. Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL).

The percent of infants being discharged from the NICU without antibiotic exposure is shown in Fig 5. At baseline, 15.8% of infants were discharged without receiving antibiotics. This number increased after implementation of the SRC and after initiating the guideline for evaluation and management of EOS in infants <35 weeks. On average, 35% of infants are now being discharged from our NICU each year without receiving antibiotics, meaning an additional 225 infants are avoiding antibiotic exposure every year.

FIGURE 5

P-chart displaying the percent of infants discharged each month without ever receiving antibiotics (total discharges = 5702). Special cause variation was noted on 2 occasions, and the center line (CL) (mean) was adjusted upward accordingly. Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL).

FIGURE 5

P-chart displaying the percent of infants discharged each month without ever receiving antibiotics (total discharges = 5702). Special cause variation was noted on 2 occasions, and the center line (CL) (mean) was adjusted upward accordingly. Dashed lines indicate the upper control limit (UCL) and lower control limit (LCL).

Process measures suggest that EOS evaluations in infants ≥36 weeks decreased by 42.3% or 140 infants per year who no longer undergo laboratory evaluations or receive antibiotics unnecessarily. Similarly, 90 fewer infants born <35 weeks are receiving empirical antibiotics each year. Compliance with other guidelines, such as duration of antibiotic use for congenital pneumonia, was high (Supplemental Fig 7). Overall, the AUR has declined by 43.8%. There were no known cases of missed bacteremia, and rates of nosocomial sepsis and infection-related mortality were unchanged (Supplemental Table 2).

We describe a comprehensive, interprofessional initiative conducted in the context of a multicenter learning collaborative that safely decreased unnecessary antibiotic use and antibiotic exposure in a level 4 NICU. Our team applied core principles of QI that included stakeholder involvement, establishing an aim statement and key driver diagram, testing changes with plan-do-study-act cycles, and measuring data over time. This robust approach allowed our NICU to make substantial gains toward attaining each of the 7 core elements of hospital antibiotic stewardship programs outlined by the Centers for Disease Control and Prevention (leadership commitment, accountability, drug expertise, action, tracking, reporting, and education).17  In doing so, we were able to improve the processes of antibiotic prescribing and use in our NICU while avoiding harm. We also believe that our comprehensive, interprofessional approach was necessary to change our unit’s culture surrounding antibiotic prescribing and assisted in achieving and surpassing our project aim.

The current observed monthly AUR of 15.5% has been sustained for >20 months, suggesting that true change has occurred in our system. Several tests of change corresponded with special cause variation observed in our AUR, and special cause variation was noted in our process measures, both of which support and strengthen our conclusion that our interventions have positively impacted our primary outcome measure. We believe the decline in the AUR was most attributable to several important interventions: implementation of the SRC, adoption of a 36-hour rule-out period for sepsis evaluations, creating a 36-hour default duration for all antibiotics, and development of guidelines for common infections. Significant opportunity to decrease unnecessary antibiotic use in the evaluation of EOS existed in our NICU. Numerous published studies have revealed a decline in empirical antibiotics for EOS with use of the SRC, and authors of a recent meta-analysis found that compared to conventional management, use of the SRC was associated with a relative risk of empirical antibiotic use of 45%.18  Our guideline for EOS in infants <35 weeks also impacted antibiotic use and antibiotic exposure and is an approach that could be applicable to most, if not all, NICUs caring for preterm infants. Moreover, our approach is consistent with the newest recommendations from the American Academy of Pediatrics Committee on Fetus and Newborn for management of infants ≤34 6/7 weeks with suspected or proven EOS.19  Although strong evidence was not always available when developing several other guidelines for infections, our approach to standardize practice through consistent definitions, choice and duration of antibiotic therapy, and diagnostic stewardship was important to reduce unwanted practice variation.

Sustained improvement, often referred to as “holding the gains,” can be challenging in complex health care systems. We believe sustained improvement was made possible by several factors. First, use of frequent audits and feedback, such as posting the AUR in the NICU and providing provider-specific AURs, helped motivate all those working in the NICU, even if they were not actively participating on the team, to achieve our goal. Second, although educating staff on risks of antibiotic overuse and misuse was necessary, we relied on changes to the system to achieve our desired outcome. Examples include the antibiotic hard stop and use of the SRC for all infants of a stated gestational age. Lastly, focusing on the Centers for Disease Control and Prevention core elements of antibiotic stewardship fostered a comprehensive approach, which ensured that all areas of significant antibiotic overuse in our NICU were addressed.

Other antibiotic stewardship efforts have been described in recent years. Many have been focused on a single disease process, such as EOS, or a single intervention, such as an automatic stop to antibiotic orders.20,21  Unlike those efforts, our initiative represents a comprehensive approach to antibiotic stewardship in the NICU using rigorous QI methods. Use of control charts to display outcome and process measures helped us understand whether changes were having the desired impact and, importantly, detect improvements in real time. Although measurement of antibiotic use is not new, our secondary outcome measure of antibiotic exposure during the NICU stay is novel. The increase in the percentage of infants discharged without receiving antibiotics, coinciding with a decline in the AUR, supports our conclusion that many infants were being exposed to antibiotics unnecessarily.

The Choosing Antibiotics Wisely iNICQ revealed significant increases in each of the 7 core elements of antibiotic stewardship programs and simultaneously observed a 34% relative risk reduction in the AUR among participating centers.22  Multicenter learning collaboratives have been an effective vehicle to support and drive improvement in neonatal care at the local level for a variety of important outcomes. Examples of achievements resulting from collaboratives include improvements in breastfeeding rates, nosocomial infections, CLABSI, admission hypothermia, postnatal growth failure, and care for newborns with narcotic abstinence syndrome.2328  Through participation in Choosing Antibiotics Wisely iNICQ, our team benefitted from access to content experts, QI experts, and the ability to collaborate with and learn from other NICUs focused on the same goal: reducing unnecessary antibiotic use.

Limitations of our initiative include that it reflects practice changes occurring in a single NICU and hospital system, some of which may not be applicable to other units or hospitals. Practice changes that impacted care provided in well-infant units, such as use of the SRC, was made easier by the fact that nearly all well newborns are staffed by a pediatric hospitalist or NICU APP, vastly decreasing our number needed to influence. In addition, the primary drivers of unnecessary antibiotic use in our NICU may not be the same in other NICUs with varying patient populations, which underscores the important role that context plays in QI.

Our initiative successfully and safely reduced unnecessary antibiotic use; all NICUs are encouraged to address antibiotic stewardship in their local context. We believe our methods for achieving improvement, which involved a rich understanding of our current system, rigorous testing, standardization of practices, and analyzing data over time, are applicable to all teams focused on QI.

We recognize the members of the Golisano Children’s Hospital NICU antibiotic stewardship team who participated in this initiative: Sanjiv Amin, MBBS, MD, MS; Meghan Baldo, PharmD, BCPPS; Claire Burke, PharmD; Patricia Chess, MD; Cynthia Christy, MD; Andrew Dylag, MD; Luke Guerrero (parent); Whitney Guerrero (parent); Sema Hart, MD; Victor Hernandez, PNP, NNP; Marilyn Hyman, BSN, RNC-NIC; Nirupama Laroia, MD; Ayesa Mian, MD; Karen Miller, RN, BSN; Molly Miller, PNP, NNP; Yukiko Miura, MD; Suzanne Mullin, MD; Jennifer Nayak, MD; Lisa Paolucci, RN, BSN; Laura Price, MD; Diane Prinzing; Gloria Pryhuber, MD; Julie Riccio, MD; Kathryn Shapiro, MS, RN, CHSE; Kristin Scheible, MD; Shanley Sifain, MD; Barbara Springer, MS, RNC-NIC; Laurie Steiner, MD; Timothy P. Stevens, MD, MPH; Robert Swantz, MD; Brenda Tesini, MD; Elaine Trottier, RN BSN; Perihan Ulema, DO; Derek Wakeman, MD; and Geoffrey A. Weinberg, MD.

Dr Meyers served as the project lead for all 3 years and drafted the initial manuscript; Ms Tulloch assisted with data collection and data preparation; Ms Brown assisted with implementation of key interventions for the initiative; Drs Caserta and D’Angio were key leaders on critical aspects of the initiative and reviewed data; and all authors reviewed and revised the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.

FUNDING: No external funding.

     
  • APP

    advanced practice provider

  •  
  • AUR

    antibiotic use rate

  •  
  • CLABSI

    central line–associated bloodstream infection

  •  
  • EMR

    electronic medical record

  •  
  • EOS

    early-onset sepsis

  •  
  • iNICQ

    Internet-based Newborn Improvement Collaborative for Quality

  •  
  • NEC

    necrotizing enterocolitis

  •  
  • QI

    quality improvement

  •  
  • SRC

    sepsis risk calculator

  •  
  • UTI

    urinary tract infection

1
Alexander
VN
,
Northrup
V
,
Bizzarro
MJ
.
Antibiotic exposure in the newborn intensive care unit and the risk of necrotizing enterocolitis
.
J Pediatr
.
2011
;
159
(
3
):
392
397
2
Kuppala
VS
,
Meinzen-Derr
J
,
Morrow
AL
,
Schibler
KR
.
Prolonged initial empirical antibiotic treatment is associated with adverse outcomes in premature infants
.
J Pediatr
.
2011
;
159
(
5
):
720
725
3
Cantey
JB
,
Pyle
AK
,
Wozniak
PS
,
Hynan
LS
,
Sánchez
PJ
.
Early antibiotic exposure and adverse outcomes in preterm, very low birth weight infants
.
J Pediatr
.
2018
;
203
:
62
67
4
Ting
JY
,
Synnes
A
,
Roberts
A
, et al;
Canadian Neonatal Network Investigators
.
Association between antibiotic use and neonatal mortality and morbidities in very low-birth-weight infants without culture-proven sepsis or necrotizing enterocolitis
.
JAMA Pediatr
.
2016
;
170
(
12
):
1181
1187
5
Cantey
JB
,
Milstone
AM
.
Bloodstream infections: epidemiology and resistance
.
Clin Perinatol
.
2015
;
42
(
1
):
1
16, vii
6
Cotten
CM
,
McDonald
S
,
Stoll
B
,
Goldberg
RN
,
Poole
K
,
Benjamin
DK
 Jr.
;
National Institute for Child Health and Human Development Neonatal Research Network
.
The association of third-generation cephalosporin use and invasive candidiasis in extremely low birth-weight infants
.
Pediatrics
.
2006
;
118
(
2
):
717
722
7
Cotten
CM
.
Adverse consequences of neonatal antibiotic exposure
.
Curr Opin Pediatr
.
2016
;
28
(
2
):
141
149
8
Ho
T
,
Buus-Frank
ME
,
Edwards
EM
, et al
.
Adherence of newborn-specific antibiotic stewardship programs to CDC recommendations
.
Pediatrics
.
2018
;
142
(
6
):
e20174322
9
Schulman
J
,
Dimand
RJ
,
Lee
HC
,
Duenas
GV
,
Bennett
MV
,
Gould
JB
.
Neonatal intensive care unit antibiotic use
.
Pediatrics
.
2015
;
135
(
5
):
826
833
10
Escobar
GJ
,
Puopolo
KM
,
Wi
S
, et al
.
Stratification of risk of early-onset sepsis in newborns ≥ 34 weeks’ gestation
.
Pediatrics
.
2014
;
133
(
1
):
30
36
11
Biondi
EA
,
Mischler
M
,
Jerardi
KE
, et al;
Pediatric Research in Inpatient Settings (PRIS) Network
.
Blood culture time to positivity in febrile infants with bacteremia
.
JAMA Pediatr
.
2014
;
168
(
9
):
844
849
12
Lefebvre
CE
,
Renaud
C
,
Chartrand
C
.
Time to positivity of blood cultures in infants 0 to 90 days old presenting to the emergency department: is 36 hours enough?
J Pediatric Infect Dis Soc
.
2017
;
6
(
1
):
28
32
13
Mukhopadhyay
S
,
Puopolo
KM
.
Clinical and microbiologic characteristics of early-onset sepsis among very low birth weight infants: opportunities for antibiotic stewardship
.
Pediatr Infect Dis J
.
2017
;
36
(
5
):
477
481
14
Puopolo
KM
,
Mukhopadhyay
S
,
Hansen
NI
, et al;
NICHD Neonatal Research Network
.
Identification of extremely premature infants at low risk for early-onset sepsis
.
Pediatrics
.
2017
;
140
(
5
):
e20170925
15
Schulman
J
,
Profit
J
,
Lee
HC
, et al
.
Variations in neonatal antibiotic use
.
Pediatrics
.
2018
;
142
(
3
):
e20180115
16
Provost
LP
,
Murray
SK
.
The Healthcare Data Guide: Learning from Data for Improvement
.
San Francisco, CA
:
Jossey-Bass
;
2011
17
Centers for Disease Control and Prevention
.
Core elements of hospital antibiotic stewardship programs.
2014
. Available at: www.cdc.gov/getsmart/healthcare/implementation/core-elements.html. Accessed December 1, 2019
18
Achten
NB
,
Klingenberg
C
,
Benitz
WE
, et al
.
Association of use of the neonatal early-onset sepsis calculator with reduction in antibiotic therapy and safety: a systematic review and meta-analysis
.
JAMA Pediatr
.
2019
;
173
(
11
):
1032
1040
19
Puopolo
KM
,
Benitz
WE
,
Zaoutis
TE
;
COMMITTEE ON FETUS AND NEWBORN
;
COMMITTEE ON INFECTIOUS DISEASES
.
Management of neonates born at ≤34 6/7 weeks’ gestation with suspected or proven early-onset bacterial sepsis
.
Pediatrics
.
2018
;
142
(
6
):
e20182896
20
Arora
V
,
Strunk
D
,
Furqan
SH
, et al
.
Optimizing antibiotic use for early onset sepsis: a tertiary NICU experience
.
J Neonatal Perinatal Med
.
2019
;
12
(
3
):
301
312
21
Astorga
MC
,
Piscitello
KJ
,
Menda
N
, et al
.
Antibiotic stewardship in the neonatal intensive care unit: effects of an automatic 48-hour antibiotic stop order on antibiotic use
.
J Pediatric Infect Dis Soc
.
2019
;
8
(
4
):
310
316
22
Dukhovny
D
,
Buus-Frank
ME
,
Edwards
EM
, et al
.
A collaborative multicenter QI initiative to improve antibiotic stewardship in newborns
.
Pediatrics
.
2019
;
144
(
6
):
e20190589
23
Billett
AL
,
Colletti
RB
,
Mandel
KE
, et al
.
Exemplar pediatric collaborative improvement networks: achieving results
.
Pediatrics
.
2013
;
131
(
suppl 4
):
S196
S203
24
Schulman
J
,
Stricof
RL
,
Stevens
TP
, et al;
New York State Regional Perinatal Centers
;
New York State Department of Health
.
Development of a statewide collaborative to decrease NICU central line-associated bloodstream infections
.
J Perinatol
.
2009
;
29
(
9
):
591
599
25
Stevens
TP
,
Shields
E
,
Campbell
D
, et al
.
Statewide initiative to reduce postnatal growth restriction among infants <31 weeks of gestation
.
J Pediatr
.
2018
;
197
:
82.e2
-
89.e2
26
Walsh
MC
,
Crowley
M
,
Wexelblatt
S
, et al;
Ohio Perinatal Quality Collaborative
.
Ohio Perinatal Quality Collaborative improves care of neonatal narcotic abstinence syndrome
.
Pediatrics
.
2018
;
141
(
4
):
e20170900
27
Ware
JL
,
Schetzina
KE
,
Morad
A
,
Barker
B
,
Scott
TA
,
Grubb
PH
.
A statewide quality improvement collaborative to increase breastfeeding rates in Tennessee
.
Breastfeed Med
.
2018
;
13
(
4
):
292
300
28
Wirtschafter
DD
,
Powers
RJ
,
Pettit
JS
, et al
.
Nosocomial infection reduction in VLBW infants with a statewide quality-improvement model
.
Pediatrics
.
2011
;
127
(
3
):
419
426

Competing Interests

POTENTIAL CONFLICTS OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose

Supplementary data