Keywords
antibiotic stewardship program - antibiotic use rate - neonatal early-onset sepsis
- neonatal intensive care unit - sepsis risk calculator
The incidence of neonatal early-onset sepsis (EOS) has declined by nearly four-fold
since guidelines from the Centers for Disease Control and Prevention (CDC) and the
American Academy of Pediatrics (AAP) recommended group B Streptococcus antibiotics
prophylaxis.[1]
[2]
[3] However, this has led to increased use of laboratory testing and antibiotics for
bloodstream infections in newborn infants, many of whom show no clinical symptoms
of sepsis. A negative blood culture is required to discontinue antibiotics after 48 hours.
Even in cases with a negative culture result, the lack of consensus regarding antibiotic
use in newborns with suspected sepsis often results in prolonged antibiotic use.[4] The fact that only 3 to 8% of those screened for EOS have culture-proven sepsis
highlights the cautious nature of conventional approaches to the management of neonatal
sepsis.[5]
[6]
The prolonged use of broad-spectrum antibiotics is not without consequences. Prolonged
exposure to antibiotics can disrupt the normal flora in the gut, leading to dysbiosis
and selection of more aggressive and pathogenic bacteria.[7] Cotten et al showed that infants receiving more than 5 days of antibiotics for EOS
had increased risk for developing necrotizing enterocolitis.[8] Indiscriminate use of antibiotics can lead to antibiotic resistance.[9]
[10] There is also concern for long-term consequences of early antibiotic exposure, including
the development of obesity later in life.[11]
[12]
[13] Antibiotic prescribing practices vary widely among neonatal intensive care units
(NICUs), from 2.4 to 97.1%.[14]
[15] However, despite the wide range of antibiotic prescribing practices, clinical outcomes
remain similar, suggesting that antibiotics are often prescribed unnecessarily.[14]
[15]
In the last few years, there has been an organized approach nationally to decrease
antibiotic exposure to help reduce the development of antibiotic resistance. While
the experience of antibiotic stewardship programs (ASPs) for adult populations have
been extensively reported,[16]
[17]
[18] there have been fewer reports in the newborn population, especially those in nonacademic
community NICUs.[19]
[20]
[21] Our NICU had been tracking antibiotic use rate (AUR) on a yearly basis for all antibiotic
usage including EOS and late-onset sepsis. AUR of 22% for all antibiotic exposure
was identified as a significant problem at the Pomona Valley Hospital Medical Center
(PVHMC) NICU in 2018. This was reviewed in the context of low incidence of EOS in
our NICU of 0.6/1,000 live births in the preceding years. We believe that lack of
standardized antibiotic prescription guidelines as well as lack of consensus on when
to stop antibiotics with negative cultures contributed to excessive use of antibiotics.
Schulman et al noted that in NICU's with ASPs, the AUR declined by 28.7% as compared
with 16.2% decline in NICU's that did not have an ASP.[15] Hence, we set out to revise our institutional methodology in ways that would decrease
antibiotic exposure for newborn infants with suspected sepsis in our NICU. We launched
this quality improvement (QI) program based on the Institute of Health Care Improvement
(IHI) model. The IHI model is a simple and effective tool used by health care organizations
to promote positive change, increase quality, and reduce cost. Our aim was to decrease
the overall AUR in NICU from 22% at baseline to 17% by the end of 2021.
Materials and Methods
PVHMC NICU is a 53-bed level 3B NICU. The NICU receives referrals from Labor and Delivery
Department with approximately 6,000 births annually. Neonatal and maternal transport
teams bring in additional referrals from the surrounding Inland Empire region that
includes the counties of Los Angeles, San Bernardino, Riverside, Mono, and Kern County.
The NICU provides care to infants with gestational age of 22 weeks and higher as well
as sick infants requiring surgical interventions, inhaled nitric oxide, and therapeutic
hypothermia. Infants with complex heart conditions and need for extracorporeal membrane
oxygenation are referred to the nearby tertiary center. Symptomatic infants with suspected
EOS are admitted to the NICU. Asymptomatic babies with history of maternal chorioamnionitis
are cared for in the newborn nursery and are not part of this stewardship program.
The Newborn Nursery is a separate administrative unit governed directly by the Department
of Pediatrics and was not included in the NICU ASP. Neonatology service is consulted
when these infants become symptomatic requiring NICU care. A small number of babies
may be treated for 48 hours in the newborn nursery.
Interventions
The interventions that comprised the ASP were first implemented in 2018 ([Table 1]). The success or failure of each intervention was critically evaluated on a quarterly
basis, and interventions were modified based on the results of analysis.
Table 1
Change implementation timeline
|
Quarter
|
Implementation
|
|
Q3 2018
|
Antibiotic stewardship team
|
|
Q1 2019
|
Sepsis calculator
Sepsis ruled out at 48 h
Antibiotic stewardship rounds
|
|
Q1 2020
|
36-h soft stop
|
|
Q3 2020
|
Electronic medical record linked to the sepsis calculator and maternal information
|
|
Q4 2020
|
36-h hard stop for all NICU antibiotic orders
|
Abbreviation: NICU, neonatal intensive care unit.
We assembled a multidisciplinary NICU antibiotic stewardship team in the third quarter
of 2018. The stewardship team included neonatologists, NICU medical director, NICU
quality supervisor, NICU nursing staff, pharmacists, Medical Director of PVHMC infection
control department, and an infection prevention nurse.
The team met weekly for 2 months and then met monthly. The team collectively reviewed
available clinical literature on EOS that included guidance from AAP. AUR data were
collected quarterly and reviewed by the stewardship team. Team members also examined
NICU antibiotic prescription practices in effect, the NICU specific antibiogram, and
antibiotic sensitivity pattern. The team visited the affiliated tertiary center to
exchange ideas with staff and learn about the practices used by the established ASP
at that institution. A key driver diagram that identifies what drives or contributes
to the project aim was developed ([Fig. 1]).
Fig. 1 Driver diagram. AAP, American Academy of Pediatrics; CDC, Centers for Disease Control
and Prevention; EMR, electronic medical record; NICU, neonatal intensive care unit;
PDSA, Plan Do Study Act; PVHMC, Pomona Valley Hospital Medical Center; VON, Vermont
Oxford Network.
In Quarter 1 (Q1) of 2019, the decision was made to update the institutional methodology
for NICU antibiotic prescription.
Review of our data led to the development of strategies for improvement that included
creation of standardized guidelines for management of suspected sepsis ([Figs. 2] and [3]),[22] periodic reviews, implementation of antibiotic stewardship rounds, use of Plan Do
Study Act (PDSA) cycles, and review of any patients that required readmission due
to sepsis-related illness. The guidelines for EOS included the use of sepsis risk
calculator (sepsiscalculator.kaiserpermanente.org). Physician use of Kaiser sepsis
risk calculator was followed on a monthly basis. Sepsis evaluation generally included
a complete blood count with manual differential count and a blood culture obtained
via either peripheral vein or an umbilical catheter when applicable.
Fig. 2 EOS guidelines for Infants 35 weeks or greater gestation. AAP, American Academy of
Pediatrics; EOS, early-onset sepsis; GBS, group B streptococcus; IPA, intrapartum
antibiotic; ROM, rupture of membrane.
Fig. 3 EOS guidelines for infants 346/7 or less gestation. AAP, American Academy of Pediatrics; CBC, complete blood count;
EOS, early-onset sepsis; GBS, group B streptococcus; IAI, intra-amniotic infection;
PROM, premature rupture of membranes; ROM, rupture of membrane.
Stewardship rounds initially included pharmacists rounding with the attending physicians.
The pharmacist collected pertinent clinical and laboratory data on a daily basis for
any baby on antibiotics, and data were reviewed with the team of physicians. A negative
blood culture at 48 hours was considered sufficient evidence for the discontinuation
of antibiotic treatment. The pharmacist assumed the responsibility of checking the
culture results at 48 hours and discontinuing the antibiotics. If antibiotics were
continued beyond 48 hours despite negative blood culture results, physicians were
instructed to document the reasoning for continued antibiotic use in the electronic
medical record (EMR). The pharmacist created a stewardship document in the EMR each
day to document the consensus reached at the stewardship rounds.
2019
Sepsis Probability Risk Calculator
Treatment of infants 35 weeks or greater was directed by the use of the EOS calculator
and documented EOS risk per 1,000 births. Chart review revealed poor physician compliance
with use and documentation of the EOS risk. In Q3 2019 physicians were asked for feedback
on the EOS calculator use. The physicians reported that navigating to another website
outside the EMR in order to use the sepsis calculator was cumbersome and time consuming.
The need for a process that provided easier access to obtain the maternal information
for the calculator was noted. In the meantime, physicians were reeducated about the
need to document an EOS risk on all infants greater than or equal to 35 weeks in the
EMR and treat accordingly. Education took place during handoff rounds, via email communications,
and during discussions at monthly NICU meetings. The Medical Director completed review
of cases with missing EOS risk documentation and provided follow up. We noted increased
use of the EOS calculator, from 51.1% in Q3 2019 to 81 to 90% by Q4 2019 ([Fig. 4]). The EOS calculator web page was linked to the EMR in Q3 2020, which improved the
overall process and efficiency for the providers. Compliance with the sepsis calculator
was noted 90% or greater in Q4 2020.
Fig. 4 2019 early-onset sepsis calculator compliance 35 weeks or greater.
Antibiotic Stewardship Rounds
The antibiotic stewardship rounds were introduced in Q1 2019. These rounds were designed
to have a daily group discussion between the pharmacist and the physicians rounding
on a particular day. In Q4 2019, a review of the stewardship rounding process revealed
that pharmacists were often rounding with individual physicians rather than the whole
team. To address this inconsistency, the stewardship team established a consistent
time and place for the rounding process. The stewardship rounds were incorporated
into the daily physician handoff that occurred every morning. All physicians that
would be working that day, as well as the physician on call the previous night, were
present for the handoff between shifts. The pharmacist that would be working that
day also participated in rounds. All infant cases scheduled to receive antibiotics
that day were discussed. Antibiotic treatment was discontinued based on the consensus
reached during these rounds. The pharmacist was responsible for daily creation of
a stewardship progress note and for placing any required orders for the discontinuation
of antibiotic treatment.
Antibiotic Dosages
In Q4 2019, review of the October data revealed that 11.3% of infants with gestational
age >35 weeks had received unintended extra doses of antibiotics. Review of these
cases with the pharmacy manager identified poor oversight and poor coverage over the
weekends as contributing factors. A core pharmacy team was developed. Weekend coverage
was improved by staffing core pharmacy team members that were knowledgeable about
the use of the sepsis risk calculator. Institutional protocol was updated to require
stewardship rounds be increased from 5 to 7 days a week. This resulted in a progressive
decrease of unintended doses administered during Q4 2019 ([Fig. 5]).
Fig. 5 Extra antibiotic dosages 35 weeks or greater 2019 to 2020.
2020
36-Hour Stop
In Q1 2020, the 48-hour antibiotic stop was transitioned to a 36-hour stop. With this
modification to institutional protocol, the pharmacist was expected to place the 36-hour
stop after stewardship rounds if it was ascertained that sepsis had been ruled out.
36-Hour Hard Stop
In Q1 2020, it was noted that the 36-hour stop was not being followed consistently.
Some physicians were uncomfortable discontinuing the antibiotics at 36 hours.
To address this concern, the hospital's infection control department was asked to
review all positive blood cultures in the NICU for the previous year. This review
concluded that in the newborn population, all positive cultures were noted to have
become positive within 36 hours. Hence, the stewardship team agreed upon a 36-hour
hard stop. In Q4 2020, education to all staff about the new process took place and
the 36-hour hard stop was implemented for all NICU antibiotics. In consultation with
the information technology department a process was implemented for automatic discontinuation
of any antibiotics prescribed in the NICU at 36 hours. If the infant required antibiotics
beyond 36 hours, a new order was required to be placed in the EMR.
The 3-day antibiotic treatment of infants 35 weeks or greater decreased from 26% in
2020 to 18% in 2021 ([Fig. 6]).
Fig. 6 Early-onset sepsis treatment days 35 weeks or greater.
External Link
To streamline the process of obtaining maternal information and the sepsis calculator
use, an external link was created in the EMR. The external link directs the physician
to maternal data needed for the sepsis calculator consolidated in one place. A link
to the Kaiser sepsis calculator was also inserted on the same page that would open
up the sepsis calculator without having to navigate out of the EMR.
Analysis
Data were obtained from the NICU EMR system. Statistical control charts (P Chart)
was used to plot the AUR quarterly. The data on sepsis-related readmissions were also
obtained from the EMR.
Ethical Considerations
The institutional review board (IRB) at PVHMC approved the study. The protocol submitted
to the IRB was updated periodically to document project progress.
Results
Primary Measures
The primary outcome measure was AUR in the NICU. AUR is calculated as a percentage
total number of NICU days of parenteral antibiotic exposure, divided by the total
number of NICU days for that year. A decrease in AUR was noticed after the antibiotic
stewardship program was implemented in July 2018 ([Fig. 7]). In the fourth quarter of 2018, while the literature was being reviewed, the physicians
became aware of the most recent guidelines for treating suspected sepsis. The practice
of antibiotic prescription began to change as newer guidelines were being made. By
the end of Q1 2019, the sepsis calculator, the 48-hour stop, and the antibiotic stewardship
rounds were formally put into practice. Further decline in AUR was noticed after these
interventions.
Fig. 7 Antibiotic use rate P chart.
After the initial interventions of sepsis calculator use, antibiotic stewardship rounds,
and 48-hour antibiotic stop, the AUR demonstrated a sustained decrease. Further interventions
including the introduction of an external link to the sepsis calculator, 36-hour stop,
and a 36-hour hard stop helped sustain the results ([Fig. 7]). The decrease in AUR was sustained through the years of stewardship program from
2019 to 2022 ([Fig. 7]). By the end of the year 2022, the decrease in AUR was sustained and low at 13.18%,
reduction of approximately 40% from the baseline before beginning the stewardship
program.
During the years of 2018 to 2022, our NICU admission rate ranged between 675 and 762
patients per year. Our inborn admission rate ranged between 11 and 13% of live births.
The NICU mortality rate ranged between 9 and 12 per 1,000 admissions. The number of
patients less than 36 weeks ranged between 285 and 386 patients per year.
The algorithm in [Fig. 3] was developed as a guideline for treatment of infants 34 weeks or less. The goal
was to standardize and decrease the length of treatment with antibiotics.
Although we did not measure compliance with this guideline, we observed a decrease
in the length of antibiotic treatment. In infants 34 weeks or less gestation, the
antibiotic treatment showed improvement over the years of ASP. Infants with no treatment
did not show any significant change (36% in 2019 vs. 33% in 2022). Infants with 2-day
treatment showed improvement from 38% in 2019 to 58% in 2022. We believe that some
of these infants were originally being treated for 3 days or more. The proportion
of infants treated for 3 days demonstrated much desired change from 21% in 2019 down
to 5% in 2022. Infants receiving 7 days or longer treatment remained nearly the same
3 versus 2% for 7-day treatment and unchanged at 2% for treatment greater than 7 days.
We believe that the lesser proportion of babies treated for 3 days resulted in decreased
length of treatment overall in this group of infants.
Secondary Measures
Sepsis-related readmissions to the NICU were utilized as a balancing measure. There
were no sepsis-related cases admitted during the study period ([Table 2]). We were limited to readmission data at PVHMC. However, since the majority of these
infants belong to PVHMC health service area, we are confident that they would be readmitted
to our facility.
Table 2
Readmissions
|
Year
|
Total admission
|
Sepsis readmission
|
Other readmission
|
|
2018
|
732
|
0
|
2 Respiratory distress
|
|
2019
|
689
|
0
|
2 CLD, 2 hyperbilirubinemia
|
|
2020
|
675
|
0
|
1 Emesis, 2 hyperbilirubinemia
|
|
2021
|
762
|
0
|
1 desaturation
1 hyperbilirubinemia
|
Abbreviation: CLD, chronic lung disease.
As an additional finding, we reviewed infants with gestational age > 35 weeks treated
with antibiotics for more than 2 days. The 3-day antibiotic treatment of infants 35
weeks or greater decreased from 26% in 2020 to 18% in 2021 ([Fig. 6]).
Discussion
Decreasing antibiotic use in the NICU is a challenging task. The contributing factors
include absence of specific guidelines to initiate antibiotics in infants with suspected
sepsis and the lack of consensus on when to discontinue antibiotics in infants with
unproven sepsis.
Antibiotic stewardship interventions so far have been mostly based on established
treatment guidelines,[23]
[24]
[25] automatic antibiotic stop date,[26]
[27]
[28] prospective audit,[19] and use of sepsis risk calculator.[29]
[30] A recent meta-analysis report showed that use of the neonatal EOS calculator is
associated with a substantial reduction in the use of empirical antibiotics for suspected
EOS.[31] Recently, more comprehensive stewardship reports have been published.[16]
[17] We devised our ASP based on the CDC guidance for hospital ASPs in 2019. We incorporated
all the core elements recommended by the CDC that included leadership commitment,
reporting, tracking, active oversight, pharmacy expertise, accountability, and education.
Our stewardship program included daily antibiotic stewardship rounds, staff education,
data review, and an overall periodic review of the program for compliance. Our approach
to management of suspected sepsis was based on the guidelines recommended by the AAP
Committee on the Fetus and Newborn 2018. We were able to gradually foster a comprehensive
and multidisciplinary approach to the use of antibiotics by initiating the ASP, which
eventually became a standard of practice in our NICU. Team compliance was achieved
via education at daily stewardship rounds, email notification, discussion during daily
work rounds, and medical director review of noncompliant cases. The role of pharmacy
department was crucial in implementing the interventions. Hence, a core team of pharmacists
was assembled at the direction of the hospital's pharmacy department. This core team
was actively involved in the development and implementation of initial measures and
interventions. The rest of the pharmacy team members that would rotate through the
NICU were also periodically educated and updated on the procedures of the program.
The pharmacy department made certain adjustments in their scheduling in order to enable
the core pharmacy team members to rotate through the NICU. Hence, no additional funding
was needed on behalf of the pharmacy. Originally adopting an aim statement, a key
driver diagram, testing changes with PDSA cycles, and measuring data, we observed
a gradual and sustained improvement in the AUR for 3 years, suggesting an actual change.
In Q2 2020, the AUR was noted to increase up to the mean. Upon review, we found multiple
culture positive cases for that quarter with complicated bowel diagnoses, congenital
syphilis, pneumonia, meningitis, and cellulitis that contributed to longer antibiotic
treatment. The most recent AUR of 13.18% in the year 2022 surpassed our initial aim
of 17%. We believe that the use of standardized guidelines for antibiotic management
and institution of daily antibiotic stewardship rounds helped us significantly in
tracking the progress as well as identification of problems on a daily basis. Variations
in the use of antibiotic management was discussed at stewardship rounds, as well as
reviewed by the medical director of the NICU.
In addition, a summary of review was presented monthly at the NICU steering committee
meetings. Implementation of various interventions including use of a sepsis risk calculator,
daily antibiotic stewardship rounds, and antibiotic stop orders and testing the change
resulted in a sustained decrease in AUR.
Antibiotic stewardship is not new, but our program employed a multidisciplinary approach,
enlisting the expertise of all stakeholders. We employed established QI processes
and methods, including the IHI model and PDSA cycles of change. A 40% reduction in
AUR in our NICU prevented unnecessary use of antibiotics in many infants. By achieving
sustained improvements in AUR beyond our stated goal, we were able to change the culture
of antibiotic use in our NICU and implement improved and standardized practices of
antibiotic management.
Limitations
In the beginning, the challenges in accessing the maternal data as well as the sepsis
calculator resulted in poor compliance with the use of sepsis calculator. Our stewardship
program was implemented only in the NICU, and hence, the antibiotic use in the newborn
nursery is not addressed in this program. Other NICUs and our larger group were not
part of this program and those other units implement their own ASP based on their
local situation and practices.
Conclusion
Our approach to decreased antibiotic use in the NICU was devised after thorough evaluation
of our practices and application of evidence-based standards of care. A comprehensive
and multidisciplinary QI-based approach to antibiotic use in our NICU significantly
reduced the AUR. We believe that all NICUs should devise their own ASP based on the
existing practices and available resources using established QI processes.
