BACKGROUND:

Premature infants have bradycardia and/or desaturation events due to apnea of prematurity that resolve as the infants mature. Despite American Academy of Pediatrics guidelines recommending a standard “event-free” period before discharge, length of observation in our Intensive Care Nursery was variable. By June 2018, for infants born <36 weeks’ gestation in the Intensive Care Nursery, we aimed to standardize time to discharge after the last documented event at 5 days, when the baseline mean was 3.6 days (range 0–6 days).

METHODS:

A quality-improvement team used the Model for Improvement. Plan-do-study-act cycles improved nursing documentation of events and standardized discharge criteria after consensus on operational definitions. The outcome measure was days to discharge after last documented event. Process measures included percentage of events documented completely and correctly in the electronic medical record. Balancing measure was length of stay after 36 weeks’ corrected gestational age. We used statistical process control.

RESULTS:

The baseline event watch ranged from 0 to 6 days. After defining significant events, documentation expectations, and consensus on a 5-day “watch” before discharge, the event watch range narrowed with a mean that shifted from 3.6 to 4.8 days on X-bar S statistical process control chart. Completeness of documentation increased from 38% to 63%, and documentation of significant events increased from 38% to 88%. Length of stay after 36 weeks’ corrected gestational age was unchanged, and nursing satisfaction improved.

CONCLUSIONS:

We found decreasing variation in the management of apnea of prematurity while simultaneously improving staff satisfaction. Next steps include revising electronic medical record flowsheets and spread to network NICUs.

Preterm infants have cardiorespiratory events for a variety of reasons. Because of their immaturity, they are at risk for apnea, bradycardia, and desaturations from central apnea, obstructive apnea, or a mix of the 2. The spectrum of presentations and severity requires assessment by the team to determine the clinical significance of these events.1,2  As preterm infants mature, events generally resolve, yet waiting for this resolution can postpone discharge. The lower the gestational age at birth, the longer it may take for cardiorespiratory events to cease. There is literature to suggest that infants born more premature might be observed “event free” for longer because extremely preterm infants can have events beyond term corrected gestational age (CGA).3,4  There are also data to suggest that either the number of days of apnea or a later gestational age of cessation may adversely impact neurodevelopment, but an earlier gestational age positively correlates with both apnea as well as other complications that impact outcomes, making the risk difficult to interpret.5,6 

The paucity of evidence with regard to what makes an event clinically significant causes management to be challenging and variable. We do know existing inter-NICU variation impacts postmenstrual age at discharge, and this variability in age at discharge may be partially explained by whether infants are diagnosed with apnea.7  The American Academy of Pediatrics (AAP) Committee on the Fetus and Newborn released a consensus statement in 2016 that summarized the available evidence and recommended that individual neonatal units create a standardized approach for monitoring these at-risk infants for a period of time before discharge.8 

In the Hospital of the University of Pennsylvania (HUP) Intensive Care Nursery (ICN), there was no standard approach to the management of apnea of prematurity. Despite the AAP consensus statement, the duration of an event watch before discharge was variable and unpredictable. Nurses documented these events in the electronic medical record (EMR) but had no guidance with regard to what types of events necessitated recording. To err on the side of caution, all events were documented, even minor, self-limiting events or events during oral feeding. Depending on the attending, the events would either be counted or disregarded.

This system culminated in generalized frustration and variability in practice. Nurses were unclear about documentation guidelines and what to tell parents about pending discharge. A nursing survey before project initiation confirmed that nurses were dissatisfied with physician management of events as well as discharge planning. Residents could not predict the plan of care until the attending made a decision during rounds. Attending physicians and fellows were bothered by the volume of events documented that did not factor into decision-making. Most importantly, families were frustrated with the lack of clarity surrounding discharge.

Through the creation of a key driver diagram, we theorized that the variation in management of apnea of prematurity was due to 2 primary drivers: inconsistency in documentation of events in the EMR by nurses and a lack of an agreement as a physician group on what a clinically significant event was and what an event watch duration should be. We therefore developed a series of interventions using the Model for Improvement to standardize documentation and establish a clinical consensus.

With this rationale in mind, our SMART AIM was as follows: by June 2018, for infants born <36 weeks’ gestation in the HUP ICN, we aimed to standardize the time to discharge after the last documented cardiorespiratory event at 5 days, when the baseline mean was 3.6 days (range of 0–6 days).

The HUP ICN is a 38-bed, level III unit with ∼700 admissions annually, including 120 very low birth weight infants and an average daily census of 34 patients. Nearly all of the infants admitted are inborn, with transfers from outside NICUs being rare. The unit is divided into 4 acute care bays and a transitional unit. Providers in the ICN include neonatal nurses, neonatology attending physicians and fellows, nurse practitioners, physician assistants, and pediatric residents from the Children’s Hospital of Philadelphia.

A significant proportion of families are only able to visit at night and on weekends. Therefore, we rely heavily on covering practitioners and nursing staff to communicate updates regarding discharge. Nurses use an EMR flowsheet to document all events, which are reported on rounds.

We began by creating a multidisciplinary quality-improvement team composed of nursing leadership, clinical nurses, nurse practitioners, a pediatric resident, neonatology fellows, and neonatal attending physicians. We disseminated a survey to the nursing staff to better understand the problem. The survey was optional and distributed via e-mail, and results were collected anonymously through Research Electronic Data Capture.9  It included 7 Likert-style questions about satisfaction and comfort with the management of events. We also solicited feedback about the EMR event flowsheet, and the same survey was repeated after implementation. We collected baseline data on the duration of days that patients were on an event watch before discharge, beginning with the introduction of a new EMR in March 2017. We also did an audit of all the events documented in the unit over a 2-week period to better understand the quality and quantity being recorded. This information was used to build a Pareto chart (Fig 1).

FIGURE 1

Missing documentation: a 2-week snapshot. This is a Pareto chart identifying commonly missing parts of documentation in the EMR apneic event flowsheet.

FIGURE 1

Missing documentation: a 2-week snapshot. This is a Pareto chart identifying commonly missing parts of documentation in the EMR apneic event flowsheet.

Close modal

Once baseline data were obtained, we performed a literature review of the available evidence for the management of apnea of prematurity. We held a Division of Neonatology Clinical Consensus Conference, at which we presented this information and agreed on a series of operational definitions, which are described below.

The definition of a clinically significant event is as follows: heart rate <80 beats per minute and/or saturation <85% for >5 seconds or apnea for >20 seconds. The duration is typically estimated by the bedside nurse. Vital sign changes during oral feeding should be documented in a separate flowsheet with feeding quality scores and should not count toward the event watch. When to provide stimulation remains at the discretion of the nurses. Additionally, if they decide to stimulate an infant, regardless of the vital signs, the event is documented. It is recommended that caffeine be discontinued at 34 weeks’ CGA and that an infant be off caffeine for at least 10 days before discharge. With a described caffeine half-life range of 62 to 112 hours,10  we felt that 5 days was a reasonable estimation of when a level would be subtherapeutic and then added the 5-day watch. For infants who correct to 34 weeks, the heart rate limit on the monitor can be decreased to 80 beats per minute to reduce unnecessary alarms. The event watch is defined as the number of days an infant is observed before discharge after the last event was documented in the EMR, and all infants are observed for a 5-day period after their last event before discharge (Table 1).

TABLE 1

Operational Definitions for Unit Protocol

CategoryOperational Definition
Clinically significant event Heart rate <80 and/or saturation <85 for >5 seconds 
 Apnea >20 seconds 
Discharge criteria Discharge 5 days after last significant event 
 Caffeine discontinued for 10 days before discharge 
Oral feeding–related events Do not necessarily preclude discharge 
Alarm parameters Heart rate limit lowered to 80 for infants corrected to 34 weeks 
CategoryOperational Definition
Clinically significant event Heart rate <80 and/or saturation <85 for >5 seconds 
 Apnea >20 seconds 
Discharge criteria Discharge 5 days after last significant event 
 Caffeine discontinued for 10 days before discharge 
Oral feeding–related events Do not necessarily preclude discharge 
Alarm parameters Heart rate limit lowered to 80 for infants corrected to 34 weeks 

Our unit uses General Electric Solar 8000i monitors and Nellcor N395 pulse oximeters with an averaging time of 4 seconds. Oxygen saturation limits are set on the basis of CGA, and heart rate lower limit is set at 100 until an infant reaches 34 weeks’ CGA.

After the clinical consensus, our quality-improvement team presented the changes in documentation and management to nurses through in-services and weekly e-mails. We then began plan-do-study-act (PDSA) cycles. Our first test of change was trialing documentation changes for nurses in the transitional unit, housing 6 to 8 infants approaching discharge. Small reminder cards with the documentation highlights were placed on the computers, and adherence was followed. We expanded to the rest of the unit, testing the changes in 1 bay at a time. We learned we needed to better clarify how and where to document events with oral feeding and also create a component of resident education. Two months after initiating the documentation changes, noting a decrease in nursing compliance, we added intermittent reminders during nursing shift-change huddles and via the weekly update e-mails.

Once the majority of events in the EMR met criteria for being clinically significant, we asked our clinicians to adhere to the consensus-driven 5-day event watch. We presented our main outcome data monthly at division morbidity and mortality conferences so that clinicians were aware of their performance and the progress of the project and could provide feedback.

Data for all measures were obtained through EMR chart review done on a biweekly basis by a nurse and physician member of the team.

The outcome measure was duration in days that a patient was on an event watch before discharge. We measured the number of days after the last event was documented in the EMR. Infants included in this outcome measure were those for whom event resolution was the only remaining discharge criteria.

The first process measure was the percentage of events documented in the EMR that were clinically significant per the definition agreed on in our clinical consensus. All events for infants of ≥34 weeks’ CGA were audited biweekly. The second process measure was the percentage of events that were documented completely. The EMR event flowsheet at our institution requests multiple pieces of information for each event, such as heart rate, saturation, duration, color change, and the presence of apnea. Completeness of documentation was also audited biweekly in infants >34 weeks’ CGA. Only events that were documented in the EMR by nurses were included in these data; we do not routinely look at monitor data retrospectively.

We did not want to inadvertently increase length of stay. Our balancing measure was length of stay in days beyond 36 weeks’ CGA for all infants who were ever treated with caffeine.

Data were interpreted by using a combination of statistical process control charts (SPCCs) and run charts. The main outcome measure of duration in days of an event watch was assessed by using an X-bar S SPCC. We initially tracked our data using an individuals-and-moving-range SPCC and converted to an X-bar S when we had enough data to group by month. We used run charts to follow our process and balancing measures to assess whether changes in these metrics were reflected in our outcome.11  All charts were annotated with PDSA cycles and interventions, and control limits were adjusted by using traditional health care special cause variation rules.12 

This project was determined to be a quality-improvement initiative by the Institutional Review Board of the University of Pennsylvania and was therefore exempted from further institutional review board review. The article was written following the Standards for Quality Improvement Reporting Excellence 2.0 guidelines.13 

Data were collected on all HUP ICN patients born <36 weeks’ gestation between March 2017 and February 2019. Our division’s Clinical Consensus Conference was held in October 2017, with subsequent interventions being implemented. See Table 2 for a project time line.

TABLE 2

Time Line of Interventions and PDSA Cycles

DatePDSA or Intervention
October 2017 Division Clinical Consensus Conference 
November 2017 Nursing in-service 
November 2017 First PDSA in transitional nursery 
December 2017 PDSAs spread to entire ICN 
January 2018 Nursing refresher 
February 2018 Clinicians adhere to 5-day event watch 
DatePDSA or Intervention
October 2017 Division Clinical Consensus Conference 
November 2017 Nursing in-service 
November 2017 First PDSA in transitional nursery 
December 2017 PDSAs spread to entire ICN 
January 2018 Nursing refresher 
February 2018 Clinicians adhere to 5-day event watch 

Each month, between 2 and 9 patients were on an event watch before discharge. The mean gestational age at birth for infants on an event watch before project implementation was 30.9 weeks (SD = 3.33) and after implementation was 31.7 weeks (SD = 2.71), which was not significantly different (P = .16). Our baseline data showed a mean event watch of 3.6 days with substantial variability, as evidenced by the wide control limits and mean SD of 1.6. After our PDSA cycles, we noted significant improvement and reached a mean event watch of 4.8 days with a decreased SD of 0.4 (Fig 2 A and B). The dominant reason the mean remains <5 days is because occasionally, events during oral feeding were erroneously documented in the apneic event flowsheet. Subsequently, physicians chose to not observe for a full 5 days.

FIGURE 2

X-bar and S SPCCs for the outcome measure of duration (in days) of event watch before discharge. A, X-bar chart of the average duration of an event watch in days and grouped by month. The mean was shifted when the special cause rule of 8 points above the centerline was met after establishing a 12-point trial mean. The narrowing of the control limits over time signifies a decrease in variation in the data. B, S chart of the SD of each sample over time. Again, the mean was shifted for the same special cause rule. The decreasing SD is also consistent with reduced variation. CL, control limit; LCL, lower control limit; UCL, upper control limit.

FIGURE 2

X-bar and S SPCCs for the outcome measure of duration (in days) of event watch before discharge. A, X-bar chart of the average duration of an event watch in days and grouped by month. The mean was shifted when the special cause rule of 8 points above the centerline was met after establishing a 12-point trial mean. The narrowing of the control limits over time signifies a decrease in variation in the data. B, S chart of the SD of each sample over time. Again, the mean was shifted for the same special cause rule. The decreasing SD is also consistent with reduced variation. CL, control limit; LCL, lower control limit; UCL, upper control limit.

Close modal

Process measures also demonstrated improvement. The percentage of all events documented that now met our operational definition of being clinically significant increased from a baseline audit of 38% to a median of 88% (Fig 3). The percentage of events documented completely increased from a baseline audit of 38% to a median of 63% (Fig 4). The less substantial improvement in completely documented events was not surprising given the known nursing dissatisfaction with the EMR flowsheet. The balancing measure of length of stay beyond 36 weeks’ CGA was unchanged with a median of 17 days (Fig 5). We theorize that length of stay beyond 36 weeks did not increase despite a longer event watch because with the new operational definition of a clinically significant event, fewer events were being documented. More specifically, fewer mild events were documented, which were previously associated with shorter event watches. Additionally, we tracked the number of days that infants were off caffeine before discharge. Our baseline was 22 days, and after our interventions, the median shifted to 30 days (Fig 6).

FIGURE 3

Run chart tracking the process measure of percentage of events documented that meet the operational definition of being significant. Center line shifted when the run chart special cause rule of 6 points above the median was met after establishing the initial 10-point baseline.

FIGURE 3

Run chart tracking the process measure of percentage of events documented that meet the operational definition of being significant. Center line shifted when the run chart special cause rule of 6 points above the median was met after establishing the initial 10-point baseline.

Close modal
FIGURE 4

Run chart tracking the process measure of percentage of events that were documented completely. Center line shifted when the run chart special cause rule of 6 points above the median was met after establishing the initial 10-point baseline.

FIGURE 4

Run chart tracking the process measure of percentage of events that were documented completely. Center line shifted when the run chart special cause rule of 6 points above the median was met after establishing the initial 10-point baseline.

Close modal
FIGURE 5

Run chart tracking the balancing measure of length of stay beyond 36 weeks’ CGA. G-tube, gastrostomy tube.

FIGURE 5

Run chart tracking the balancing measure of length of stay beyond 36 weeks’ CGA. G-tube, gastrostomy tube.

Close modal
FIGURE 6

Run chart tracking the number of days off caffeine before discharge.

FIGURE 6

Run chart tracking the number of days off caffeine before discharge.

Close modal

The initial survey to nursing staff was in September 2017; it was repeated after implementation in June 2018. The response rate for both surveys was ∼40%. We demonstrated improvement in nursing satisfaction in all areas (Fig 7). In free-text comments, nurses most often noted aspects of the EMR flowsheet that they remained unhappy with and discomfort with documenting events during oral feeding elsewhere.

FIGURE 7

Select nursing survey responses. The blue bars signify responses before protocol implementation, and the orange bars signify responses after implementation.

FIGURE 7

Select nursing survey responses. The blue bars signify responses before protocol implementation, and the orange bars signify responses after implementation.

Close modal

The baseline duration that infants were observed on an event watch before discharge was widely variable. This variation was due to a lack of consensus on the management of apnea of prematurity and unpredictable documentation of events. Coming to a consensus on clinically significant events as well as how we determined discharge timing has substantially reduced that variation.

The success of our interventions is supported by improvements in the completeness of documentation as well as the percentage of time when the events that are documented are clinically significant by our new criteria. The improvement in our process measures highlights the power of our key driver diagram in identifying the root of the problem and speaks to the buy-in we had from our stakeholders. The nurses welcomed guidance with regard to what they should be documenting and now feel more empowered to talk with families about timing of discharge. Finally, with 19 months of data since the project started, we have not seen a change in length of stay beyond 36 weeks’ CGA (our balancing measure).

The concept of standardization in medicine is not new, but it remains controversial. It is cited as an effective method to improve quality and reduce cost; however, it requires buy-in because it is often paired with a perceived decrease in physician autonomy.14  The fields of pediatrics and neonatology are making strides in protocol development and the application of evidence-based medicine. In fact, other groups have published on methods for reducing variation in the management of apnea of prematurity with encouraging results.

Butler et al15  performed a prospective cohort study in a 17-bed, level 2 unit where they instituted staff education and a protocol for documenting and managing apnea. They showed decreased length of stay and reduced hospital costs in the group receiving standardized management and saw no readmissions in the intervention group. This makes sense because previous work has also shown a decline in cost-effectiveness in predischarge monitoring for apnea of prematurity as the duration of monitoring increased.16 

After noting their high rate of home monitor use, Powell et al17  sought to standardize their definitions and practices. They defined a “clinically significant cardiopulmonary event” as well as guidelines for caffeine use. They saw a dramatic decrease in caffeine and home monitor use at discharge. A slightly different approach was taken by Chandrasekharan et al,18  who created an algorithm for duration of observation after an apneic event that varied on the basis of event severity. They did not see a decrease in length of stay but did see a significant decrease in post-NICU discharge readmissions. It is worth noting that the interval of observation varies widely. In the study by Butler et al,15  the observation period is 5 days or 7 days, depending on the timing of caffeine discontinuation; in the study by Chandrasekahran et al,18  it is either 3, 5, or 7 days, depending on the severity of the event and discontinuation of caffeine. Our clinician group discussed observing infants born <27 weeks’ gestation for 7 event-free days before discharge, which does have some support from retrospectively collected data4 ; however, when we suggested this, we met resistance from some physicians and went with a uniform 5-day watch.

Although our study is unique in that it was explicitly designed as a quality-improvement project at the outset, all of these studies strive to reduce variation and show improvement in their unique outcome measures. This further supports the AAP recommendation of creating a unit-specific protocol.8 

We recognize that there are pros and cons to the specific protocol we created, in part because of the paucity of prospective data in this area. We chose a relatively conservative definition of a significant event. It will likely still encompass some brief, self-resolved events, but we made a decision that achieved buy-in from nurses and physicians. We believe our project has had a positive impact on staff and families, most notably increased nursing satisfaction. System impacts include far fewer events being documented in the EMR and presumably fewer alarms with the lower heart rate limit (although we did not measure this). Nurses are also now able to independently communicate the plan of care to families overnight.

One persistent challenge we face is how to best manage events occurring during oral feeding. Although our protocol states that these events should not be considered as contributing to an apnea event watch (consistent with AAP consensus recommendations), they remain important to consider in the context of determining an infant’s maturity. Many nurses in our unit feel that these events may not be getting the attention they merit. Drawing a clear line has been difficult and warrants further discussion. When we redesign our EMR flowsheet, we will work with nurses to create a more visible space for these events.

Because this is a quality-improvement project, there are limitations to the generalizability of the work. That being said, we believe that operational definitions are a useful place to start regardless of what other drivers exist. Another limitation was the lack of family member input. Although correlating our interventions with parental stress or anxiety would have been interesting, it would be difficult to understand the impact because no family experienced the old and new practice. Our data also only included events that were documented by nurses in the EMR because we do not routinely look back at saved monitor data. It is therefore possible that events occurred but did not get documented and thus are not included in our data. Finally, we recognize that incidence of readmission would have been a valuable balancing measure to follow. Unfortunately, infants discharged from our ICN do not remain in our primarily adult-focused hospital system and often have different surnames after discharge. For these reasons, readmissions are difficult to track.

Our internal validity may have been limited by the method with which the duration of events was measured: we asked our nurses to estimate duration. Our monitors have a 4-second averaging time, and we recognize that this, and variables such as where the nurse was when an alarm went off, makes precise estimation challenging. Ultimately, this was the most straightforward option in our system, and we valued practicality.

We have demonstrated sustainability, and the project has now moved into a spread phase because we are integrating similar protocols at other sites in the Children’s Hospital of Philadelphia Newborn Care Network. Our next steps involve revising the EMR event flowsheet to reflect the data we care most about clinically and continuing the spread. Further studies could evaluate the impact on length of stay, the correlation of documented events with monitor data, reimbursement, and outcomes after discharge, such as readmission. We successfully created and implemented a protocol for managing apnea of prematurity and subsequently reduced practice variation among providers while improving nursing satisfaction. We hope our work can benefit other units looking to develop a protocol.

Dr Coughlin developed the quality-improvement project, contributed to the organization and assessment of the project, collected data, synthesized the data for publication, and drafted and critically reviewed the manuscript; Dr Posencheg contributed to the design of the initial project, contributed to data analysis, and provided edits to all drafts of the manuscript; Ms Orfe and Ms Zachritz are key members of the quality-improvement team that contributed to the implementation of the project and data acquisition and reviewed the manuscript; Drs Meadow and Yang contributed to the organization and assessment of the quality-improvement project and reviewed and critically revised the manuscript; Dr Christ contributed to the organization and assessment of the quality-improvement project, supervised the implementation of the initiative, and provided edits to the manuscript; and all authors approved the final manuscript as submitted.

FUNDING: No external funding.

AAP

American Academy of Pediatrics

CGA

corrected gestational age

EMR

electronic medical record

HUP

Hospital of the University of Pennsylvania

ICN

Intensive Care Nursery

PDSA

plan-study-do-act

SPCC

statistical process control chart

1
Lee
H
,
Rusin
CG
,
Lake
DE
, et al
.
A new algorithm for detecting central apnea in neonates
.
Physiol Meas
.
2012
;
33
(
1
):
1
17
2
Finer
NN
,
Higgins
R
,
Kattwinkel
J
,
Martin
RJ
.
Summary proceedings from the apnea-of-prematurity group
.
Pediatrics
.
2006
;
117
(
3, pt 2
):
S47
S51
3
Merritt
TA
,
Pillers
D
,
Prows
SL
.
Early NICU discharge of very low birth weight infants: a critical review and analysis
.
Semin Neonatol
.
2003
;
8
(
2
):
95
115
4
Lorch
SA
,
Srinivasan
L
,
Escobar
GJ
.
Epidemiology of apnea and bradycardia resolution in premature infants
.
Pediatrics
.
2011
;
128
(
2
).
5
Janvier
A
,
Khairy
M
,
Kokkotis
A
,
Cormier
C
,
Messmer
D
,
Barrington
KJ
.
Apnea is associated with neurodevelopmental impairment in very low birth weight infants
.
J Perinatol
.
2004
;
24
(
12
):
763
768
6
Pillekamp
F
,
Hermann
C
,
Keller
T
,
von Gontard
A
,
Kribs
A
,
Roth
B
.
Factors influencing apnea and bradycardia of prematurity - implications for neurodevelopment
.
Neonatology
.
2007
;
91
(
3
):
155
161
7
Eichenwald
EC
,
Zupancic
JA
,
Mao
WY
,
Richardson
DK
,
McCormick
MC
,
Escobar
GJ
.
Variation in diagnosis of apnea in moderately preterm infants predicts length of stay
.
Pediatrics
.
2011
;
127
(
1
).
8
Eichenwald
EC
;
Committee on the Fetus and Newborn, American Academy of Pediatrics
.
Apnea of prematurity
.
Pediatrics
.
2016
;
137
(
1
):
e20153757
9
Harris
PA
,
Taylor
R
,
Thielke
R
,
Payne
J
,
Gonzalez
N
,
Conde
JG
.
Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support
.
J Biomed Inform
.
2009
;
42
(
2
):
377
381
10
Doyle
J
,
Davidson
D
,
Katz
S
,
Varela
M
,
Demeglio
D
,
DeCristofaro
J
.
Apnea of prematurity and caffeine pharmacokinetics: potential impact on hospital discharge
.
J Perinatol
.
2016
;
36
(
2
):
141
144
11
Provost
LP
,
Murray
S
.
The Health Care Data Guide: Learning from Data for Improvement
.
Hoboken, NJ
:
Wiley
;
2011
12
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
13
Ogrinc
G
,
Davies
L
,
Goodman
D
,
Batalden
P
,
Davidoff
F
,
Stevens
D
.
SQUIRE 2.0 (Standards for Quality Improvement Reporting Excellence): revised publication guidelines from a detailed consensus process
.
BMJ Qual Saf
.
2016
;
25
(
12
):
986
992
14
Mathews
SC
,
Pronovost
PJ
.
Physician autonomy and informed decision making: finding the balance for patient safety and quality
.
JAMA
.
2008
;
300
(
24
):
2913
2915
15
Butler
TJ
,
Firestone
KS
,
Grow
JL
,
Kantak
AD
.
Standardizing documentation and the clinical approach to apnea of prematurity reduces length of stay, improves staff satisfaction, and decreases hospital cost
.
Jt Comm J Qual Patient Saf
.
2014
;
40
(
6
):
263
269
16
Zupancic
JA
,
Richardson
DK
,
O’Brien
BJ
,
Eichenwald
EC
,
Weinstein
MC
.
Cost-effectiveness analysis of predischarge monitoring for apnea of prematurity
.
Pediatrics
.
2003
;
111
(
1
):
146
152
17
Powell
MB
,
Ahlers-Schmidt
CR
,
Engel
M
,
Bloom
BT
.
Clinically significant cardiopulmonary events and the effect of definition standardization on apnea of prematurity management
.
J Perinatol
.
2017
;
37
(
1
):
88
90
18
Chandrasekharan
P
,
Rawat
M
,
Reynolds
AM
,
Phillips
K
,
Lakshminrusimha
S
.
Apnea, bradycardia and desaturation spells in premature infants: impact of a protocol for the duration of ‘spell-free’ observation on interprovider variability and readmission rates
.
J Perinatol
.
2018
;
38
(
1
):
86
91

Competing Interests

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

FINANCIAL DISCLOSURE: The authors have indicated they have no potential conflicts of interest to disclose.