Issue 2 • Volume 6

Individual QI projects from single institutions

Stewardship Intervention to Optimize Central

Venous Catheter Utilization in Critically Ill
Children
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn

Jennifer A. Blumenthal, MD*†‡; Jennifer A. Ormsby, BSN, RN, CPN, CIC§; Dimple Mirchandani, MPH*;
Chonel A. Petti, MPH, CPHQ*; Jane Carpenter, MPH§; Maggie Geller, RN, CPHQ*;
Stephanie N. Harding, BS*; Mary O’Brien, RN, MHA/MSN, CRRN*; Thomas J. Sandora, MD, MPH†‡§;
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

Monica E. Kleinman, MD*‡; Gregory P. Priebe, MD*†‡; Nilesh M. Mehta, MD*‡

Abstract
Introduction: We aimed to describe utilization and indication(s) for long-term central venous catheters (CVCs) in a pediatric inten-
sive care unit (PICU) and identify potential strategies to decrease CVC utilization. Methods: We conducted a single-center prospec-
tive quality improvement initiative at a 30-bed PICU in a large, freestanding, academic children’s hospital. We created an electronic
report to identify patients with an indwelling CVC for 7 days and older (defined as long term). We discussed the ongoing need for
each long-term CVC with PICU clinicians at weekly interdisciplinary structured “CVC stewardship rounds.” We then made recom-
mendations around expedited removal of CVCs. We conducted multiple Plan-Do-Study-Act cycles to categorize CVC indications,
identify modifiable factors, and educate PICU clinicians. We hypothesized that CVC stewardship rounds would decrease long-term
CVC utilization in our PICU. Results: From October 2016 to September 2017, 607 long-term CVCs were eligible for the steward-
ship intervention. Compared to the preintervention period, we recorded a significant increase in peripherally inserted central cathe-
ters and a decrease in nontunneled CVCs (P < 0.001). Most patients had single- or double-lumen CVCs in both the preintervention
and intervention periods (86% and 91%, respectively). The utilization of overall long-term CVC devices, and those with modifiable
indications, decreased during the intervention period. Conclusions: A single-center QI intervention focused on PICU CVC stew-
ardship was associated with a decrease in CVC utilization. (Pediatr Qual Saf 2021;6:e389; doi: 10.1097/pq9.0000000000000389;
Published online February 12, 2021.)

INTRODUCTION increased morbidity, length of stay, and hos-

Central line-associated bloodstream infec- pital costs.1,2 The incidence of CLABSIs and
tion (CLABSI) is one of the leading health- their impact on patients remain significant
care-associated infections in the intensive despite advances in infection prevention
care unit (ICU). It is associated with strategies.3,4 Bundled insertion and main-
tenance plans improve CLABSI rates.5,6
From the *Department of Anesthesiology, Critical Care The Centers for Disease Control and
and Pain Medicine, Division of Critical Care Medicine, Prevention (CDC) guidelines for best prac-
Boston Children’s Hospital, Boston, Mass.; †Department of
Pediatrics, Division of Infectious Diseases, Boston Children’s
tices provide recommendations for insertion
Hospital, Boston, Mass.; ‡Harvard Medical School, Boston, site disinfection, dressing types and frequency,
Mass.; and §Infection Prevention and Control, Boston Children’s tubing change, and antimicrobial-coated cathe-
Hospital, Boston, Mass.
ters.7 Also, the guidelines review educational modeling,
G.P.P. and N.M.M. contributed equally to this work as co-senior authors.
which serves as the basis of CLABSI prevention.
*Corresponding author. Address: Nilesh M. Mehta, MD, Division of Critical Care
Medicine, Department of Anesthesiology, Critical Care and Pain Medicine, Boston
Pediatric-specific guidelines for central venous catheter
Children’s Hospital, Boston, MA 02115 (CVC) insertion and maintenance are based on scarce sci-
PH: 617-355-7327; Fax: 617-730-0453 entific literature than adult guidelines. Pediatric ICU cli-
Email: nilesh.mehta@childrens.harvard.edu
nicians have developed consensus-based bundles to lower
Copyright © 2021 the Author(s). Published by Wolters Kluwer Health, Inc.
This is an open access article distributed under the Creative Commons
infection rates in their population and meet national
Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and patient safety goals set by the Joint Commission.8
reproduction in any medium, provided the original work is properly cited. Interventions aimed at assessing the ongoing need for
To cite: Blumenthal JA, Ormsby JA, Mirchandani D, Petti CA, Carpenter J, Geller CVCs in pediatric intensive care unit (PICU) patients,
M, Harding SN, O’Brien M, Sandora TJ, Kleinman ME, Priebe GP, Mehta NM.
Stewardship Intervention to Optimize Central Venous Catheter Utilization in
specifically long-term CVCs, have not been adequately
Critically Ill Children. Pediatr Qual Saf 2021;6:e389. described in the literature. Long-term CVCs (>7 d dwell
Received for publication October 1, 2019; Accepted September 27, 2020. time) are at a significantly higher risk for infectious and
Published online February 12, 2021. noninfectious morbidities than short-term CVCs.5,7,9,10
DOI: 10.1097/pq9.0000000000000389 Additionally, a CLABSI incidence of 26.2% for idle CVCs

1
 Impact of a Stewardship Intervention on Central Venous Catheter Utilization Pediatric Quality and Safety

(not medically indicated) has been described in the transi- 2. just-in-time data to guide discussion and decisions
tion from the adult ICU to the inpatient wards.11 Quality around ongoing CVC use;
improvement (QI) efforts aimed at reduction of CVC days 3. one-on-one discussions with providers, including
by eliminating idle CVCs and minimizing modifiable indi- attending physicians, to understand barriers to best
cations are desirable.4,7 The discussion of CVC need, with practice in individual cases; and
an effort to encourage thoughtful use and timely removal, 4. modifying subsequent cycles based on the informa-
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn

are critical elements of “CVC stewardship.” tion and evaluating each cycle’s cumulative impact
In our PICU in 2015, CVC utilization rates were higher on CVC practice.
than the 90th percentile compared with those in other
Cycle 1: Our first PDSA cycle aimed to raise aware-
centers participating in the National Healthcare Safety
ness and educate providers on the prevalence and risks
Network (NHSN).3 Despite multipronged approaches
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

of long-term CVC use in our PICU. With informatics

targeted at improving compliance for insertion and main-
support, we created an automated report that listed all
tenance bundles, our center’s mean for CLABSI rates
patients with long-term CVCs (14 d or longer)13 and the
remained unchanged around 1.96 CLABSI/1,000 CVC
number of days the CVCs had been in place. We sent this
days. A different approach, targeted at optimizing CVC
report to attending physicians at the start of their week
indications and decreasing CVC indwelling time, seemed
of service to increase awareness of long-term CVCs. To
necessary. We aimed to establish a novel CVC stewardship
maximize the detection of eligible CVCs, we subsequently
intervention to accomplish 2 aims: (1) to understand cur-
amended the definition to include all CVCs that had been
rent CVC utilization and identify potentially modifiable
in place for 7 days or longer. With the PICU and QI com-
practices, and (2) to decrease unnecessary/modifiable CVC
mittees’ leadership support, we started our CVC steward-
use and reduce total CVC days. We hypothesized that there
ship rounds with a small group of attending providers.
would be a decrease in CVC utilization (CVC days/patient
Cycle 2: The second PDSA cycle involved discussing
days) after implementing the stewardship intervention.
each long-term CVC (7 d or longer) with the patient’s
attending provider on a subset of teams and collecting
METHODS reasons for ongoing CVC use. Options for CVC indi-
cations included high patient acuity as defined by the
The project was a single-center QI initiative. We de-iden-
provider, multiple sedative infusions (such as analgesia/
tified patient-specific data and described summary/group
sedative medications), need for parenteral nutrition, dif-
results as trends. According to institutional policies, this
ficulty obtaining peripheral venous access, or other as
project was exempt from IRB review.
described by the clinician. We categorized the indications
Boston Children’s Hospital is a 404-bed pediatric aca-
for CVC usage as modifiable or nonmodifiable a priori. A
demic institution that is Magnet designated. The 30-bed
modifiable reason included any case where an alternative
pediatric Medical–Surgical ICU (the “PICU”) has over 2,000
to CVC was deemed feasible by the stewardship group.
admissions annually. Major diagnostic categories include
We also documented if the CVC was no longer needed,
neurology/neurosurgery, oncology, bone marrow transplant,
and the treating team intended to remove the line.
and complex medical and surgical diagnoses. The PICU QI
Cycle 3: As part of this PDSA cycle, we implemented
committee consists of a QI director, consultant, and champi-
the “CVC stewardship” rounding with all providers in
ons representing physicians, nursing, respiratory therapists,
October 2016. During weekly QI rounds, the QI team
pharmacists, trainees, and administrative staff. Based on the
discussed each long-term CVC with the patient’s PICU
prevailing rates of CLABSI in the PICU, the QI committee
attending (or designee if the attending was unavailable).
determined the need for an innovative approach to reduce
The attending would select the most appropriate indica-
the CLABSI rate. The multidisciplinary committee identified
tion(s) for the CVC in that patient. We documented when
this intervention as a priority effort and included it in the
a CVC was deemed unnecessary by the attending and fol-
PICU quality management plan. A subset of the PICU QI
lowed up to provide a reminder in 72 hours if necessary.
Committee conducts weekly QI walk rounds to engage with
Cycle 4: Our fourth PDSA cycle involved further refine-
frontline providers using a structured process. The “CVC
ment of our indications into more specific categories and
stewardship rounds” described in this article are now part
labeling the indications as either potentially modifiable or
of the weekly QI walk rounds.
nonmodifiable on the day of discussion. We challenged
We employed the Model of Improvement Plan-Do-
the notion of definite or nonmodifiable indications for
Study-Act (PDSA) cycle to test small changes in short
CVC during this cycle and used discussions to educate
cycles.12 Figure 1 outlines our 5 PDSA cycles (called
and support providers for alternatives to CVC when
“cycles” for brevity).
deemed available. Nonmodifiable indications included:
The main focus of the project included:
“acute” patient, multiple continuous infusions for anal-
1. six months of increasing awareness for providers gesia/sedative medications, parenteral nutrition (PN),
using data to demonstrate the current practice and long-term chemotherapy, dialysis/pheresis, and others. In
discuss modifiable practice; these situations, removing the CVC would result in the

2
 Blumenthal et al • Pediatric Quality and Safety (2021) 6:2;e389 www.pqs.com
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

Fig. 1. PDSA cycles overview for the stewardship project.

patient’s inability to receive essential care. Patients with a We enhanced buy-in from providers by ensuring the
potentially modifiable indication included those in whom participation of unit leadership in CVC stewardship
CVC removal was planned, the need for antibiotic admin- rounds. We provided succinct and efficient information
istration, the patient awaiting a planned procedure, and and adapted rapidly to requests or suggestions for team
provider preference. The stakeholders’ consensus was members’ workflow improvement. We added a report of
that these patients could receive essential care with a the number of CVC days and the number of CVC entries
peripheral intravenous (IV) catheter. We deemed difficulty per day for each long-term CVC in the fifth PDSA cycle
obtaining peripheral venous access a nonmodifiable indi- based on feedback. The weekly CVC stewardship rounds
cation only after exploring the help from the institutional lasted between 25 and 30 minutes by the end of cycle
IV placement team. If the CVC indication was potentially five.
modifiable, the QI team discussed possible options that We analyzed the data using Microsoft Excel, SQC
would enable the provider to remove the CVC. Pack QI software (version number 7.0.18187.5; Dayton,
Cycle 5: During the fifth PDSA cycle, we added a list Ohio), and GraphPad Prism (version number 8.2.1; San
of active medications for each patient to provide a mem- Diego, Calif.). The primary outcome measures included
ory tool for improving the completeness of the indica- utilization rates of long-term CVCs (defined as the num-
tions in our documentation. We shared this list with the ber of CVCs with 7 days or longer dwell time divided by
care team during the weekly CVC stewardship rounds. the number of patient days) and modifiable CVCs during
We also added a report indicating the total number of the intervention period. We also assessed weekly trends
times per 24-hour period that each CVC was accessed. in the utilization rates for all CVCs during the interven-
We included this report to improve general awareness tion period compared to the preperiod. We employed the
both of lines that were being infrequently accessed (and CDC/NHSN definition of CVC utilization. The numer-
therefore potentially could be removed) and frequently ator of each point in Figure 2 is the number of patients
accessed lines to assure each time was necessary (given with one or more long-term CVCs (or modifiable CVCs)
the risk of introduction of bacteria associated with each on the day of rounds, and the denominator is the number
access) and to evaluate potential medication changes of patients in the unit on that day. In Figure 3, these are
from IV to enteral. grouped monthly for the long-term and modifiable CVCs.

3
 Impact of a Stewardship Intervention on Central Venous Catheter Utilization Pediatric Quality and Safety

In contrast, the “All CVCs” label in Figure 3 refers to defined by the QI committee and infection control team.
the monthly CVC days/patient days calculated by NHSN We created and analyzed statistical process control charts
definitions where every day of the month is included. for the 2 outcome measures to assess our CVC steward-
Modifiable indications for CVC continuation were a priori ship intervention’s effects over time.14 We examined these
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

Fig. 2. Process control charts showing trends in CVC utilization in the PICU.

4
 Blumenthal et al • Pediatric Quality and Safety (2021) 6:2;e389 www.pqs.com
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

Fig. 3. Temporal trends of the utilization of CVCs in the PICU. A, all CVCs. B, Long-term CVCs evaluated during the stewardship
intervention. C, Long-term CVCs with a modifiable indication.

data for any evidence of special cause variation, using the year of CVC stewardship intervention, 607 CVCs
standard definitions. We used linear regression (GraphPad met the inclusion criteria, compared to the 556, which
Prism) to determine whether the slopes of the trendlines would have been eligible for intervention in the year
of CVC utilization rates were significantly different from prior (preintervention). Diagnostic categories included
zero. Our process measure was the percent of long-term 31.4% surgical patients in the preperiod and 22.9% in
CVCs reviewed per week; this was 100% on each of the the intervention period, and 41.3% neurology/neuro-
50/52 weeks of conducted CVC stewardship rounds. surgery patients in the preperiod versus 42.2% in the
To compare CVC characteristics, we completed a chart intervention period. There were no statistically signifi-
review of all long-term CVCs for the 12 months, October cant differences between the median ages, PIM3 scores,
2015 to September 2016 (the “preperiod”). We used the mortality, or diagnostic categories between the 2 groups.
chi-squared test and Fisher’s exact test to determine the Median ICU length of stay was also similar between the
statistical difference between these characteristics for the 2 groups.
2 periods. Patient characteristic data for both the prepe- Table 2 describes the CVC characteristics for the pre-
riod and the intervention period were collected to evaluate intervention period (based on the chart review) and the
potential confounders or unintended consequences. Patient intervention period. Peripherally inserted CVCs (PICCs)
characteristics included mortality rates, Pediatric Index of made up the majority of the CVCs in each period, which
Mortality 3 (PIM3) severity scores, and ICU length of stay. increased from the preperiod to the intervention period
from 54% to 67% (P < 0.001). In comparison, tempo-
rary nontunneled CVCs decreased from 27% in the pre-
RESULTS intervention period to 13% in the intervention period
Table 1 details demographic information from the prepe- (P < 0.001). Most patients had single- or double-lumen
riod and CVC stewardship intervention period. Over CVCs in both the pre and intervention periods (86% and

5
 Impact of a Stewardship Intervention on Central Venous Catheter Utilization Pediatric Quality and Safety

Table 1. Demographic and Clinical Characteristics of increased in the “infusion requirement” category from
Patients with One or More CVC for 7 Days or Longer in a 27% to 36% (P < 0.002). Also, the proportion of CVCs
Pediatric Medical–Surgical ICU retained for long-term chemotherapy decreased from the
Pre-CVC Stewardship CVC Stewardship preperiod (30%) to the intervention period (20%). As
Evaluation Period Evaluation Period shown in Table 3, 88.5% of the long-term CVC indica-
October 1, October 1,
2015–September 30, 2016–September tions were nonmodifiable, whereas 11.5% were poten-
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn

2016 30, 2017 tially modifiable.

No. unique patients Figure 2A and B shows the process control charts for
qualifying for No. unique patients weekly utilization of long-term CVCs and CVCs deemed
Characteristic evaluation (n = 223) evaluated (n = 192)
Median age in years 3.7 (0–30.1) 2.4 (0–28.3) modifiable starting in PDSA cycle 3 when modifiable
(range) CVCs were fully defined. The weekly control charts show
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

Male sex 110 (49.3%) 115 (59.9%) that the centerline for long-term CVC utilization shifted
Average PIM3 generated 1.3% (0.3–5.5%) 1.5% (0.4–6.3%)
probability of death downward early in PDSA cycle 4 (from 0.43 to 0.35 CVC
(range) days/patient days, a 19% relative decrease). The center-
No. CLABSI 11 11
Dx Surgical 70 (31.4%) 44 (22.9%) line for modifiable CVCs shifted downward 2 months
Medical 18 (8.1%) 20 (10.4%) later, in the middle of PDSA cycle 4 (from 0.08 to 0.03
Neuro (surgical/ 92 (41.3%) 81 (42.2%)
medical)
CVC days/patient days, a > 50% relative decrease). Both
Oncology 25 (11.2%) 20 (10.4%) of the new centerlines were maintained at their lower lev-
HSCT 18 (8.1%) 27 (14.1%) els through the rest of the intervention period. However,
ICU LOS Median (range) 21 (0.2–253.7) 17.5 (0.7–205)
(d) the modifiable CVC utilization rate had a single point
Mortality 13 (5.8%) 16 (8.3%) above the upper control limit early in PDSA cycle 5 that
Dx, diagnostic category; HSCT, hematopoietic stem cell transplant. was of a unclear cause. Figure 3 compares the trendlines
of monthly CVC utilization for long-term CVCs, long-
term CVCs with modifiable indications, and all CVCs
91%, respectively). The indication for insertion, which is in the PICU (the typical NHSN-reportable metric) start-
selected by the provider at the time of insertion, was sim- ing in PDSA cycle 3. All 3 trendlines show statistically
ilar for “current access unable to support treatment” but significant downward trends in CVC utilization during

Table 2. CVC Characteristics in Patients with One or More CVC for 7 Days or Longer
Preperiod Intervention
CVC Characteristic (n = 556) Period (n = 607) P
PICC 298 (54%) 404 (67%) <0.001
Tunneled, cuffed device 66 (12%) 103 (17%) 0.014
Temporary, nontunneled CVC 150 (27%) 76 (13%) <0.001
Umbilical catheter 4 (<1%) 1 (<1%) 0.2
CVC types Hemodialysis/pheresis 37 (7%) 23 (4%) 0.027
CVC vein type Internal jugular/SVC 117 (21%) 80 (13%) <0.001
Femoral 61 (11%) 43 (7%) 0.020
Subclavian 67 (12%) 72 (12%) 0.921
Peripherally inserted 305 (55%) 411 (68%) <0.001
Other (including umbilical and portal) 6 (1%) 1 (<1%) 0.06
No. CVC Lumens One 73 (13%) 110 (18%) 0.020
Two 407 (73%) 445 (73%) 0.966
Three 76 (14%) 52 (9%) 0.005
CVC insertion location All intensive care units 203 (37%) 167 290 (48%) <0.001
MSICU (30%) 268 (44%) <0.001
MICU 11 (2%) 7 (1%) 0.255
NICU 10 (2%) 7 (1%) 0.340
CICU 15 (3%) 8 (1%) 0.091
OR/procedural unit 277 (50%) 190 (31%) <0.001
Outside hospital 10 (2%) 71 (12%) <0.001
Interventional radiology 58 (10%) 44 (7%) 0.055
General floors 8 (1%) 12 (2%) 0.481
Indications for CVC insertion* Current access unable to support treatment, or 347 (33%) 363 (33%) 0.9
lack of peripheral veins to support treatment
Dialysis/pheresis 27 (3%) 32 (3%) 0.61
Hemodynamic monitoring 49 (5%) 31 (3%) 0.03
Home parenteral therapy 16 (2%) 34 (3%) 0.02
Infusion requirement (vasoactive infusion, 288 (27%) 392 (36%) <0.002
infusion hyperosmolality, pH < 5 or > 9)
Long-term chemotherapy 311 (30%) 216 (20%) <0.001
Solid organ transplant 1 (<1%) 4 (<1%) 0.38
Stem cell transplant 13 (1%) 16 (2%) 0.37
*Denominator for this section is total indications. Many patients have multiple indications for insertion documented, all included.
CICU, cardiac intensive care unit; MICU, medical intensive care unit; MSICU, medical surgical intensive care unit; NICU, neonatal intensive care unit;
OR, operating room; SVC, superior vena cava.

6
 Blumenthal et al • Pediatric Quality and Safety (2021) 6:2;e389 www.pqs.com

Table 3. Provider-specified Indications for Ongoing CVC studies focus on daily reminders of the need for a CVC
Nonmodifiable 537 (88.5%) and use various approaches for delivering reminders,
Acute patient 192 (31.6%) including multidisciplinary teams. All of these studies
PN 104 (17.1%)
Sedation 73 (12%) noted improvement in CVC days and statistically signif-
Multiple continuous infusions 72 (11.8%) icant decreases in CLABSI rates.17–19 To our knowledge,
Long-term chemotherapy 45 (7.4%)
Sedation and PN 31 (5.1%) data on such interventions in pediatric institutions are
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn

Dialysis/pheresis 13 (2.1%) not available.

Other 7 (1.1%) The number of long-term CVCs in our PICU is high,
Potentially modifiable 70 (11.5%)
None: planned removal 28 (4.6%) and the indications for ongoing use are usually related
Difficult access 20 (3.3%) to sedation, nutrition, or patient acuity. Given the overall
Antibiotics 10 (1.6%)
Provider preference 6 (1%) decrease in the CVCs with potentially modifiable indi-
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

Preprocedure 6 (1%) cations, especially in combination with the unchanging

total CVCs over time, we believe our effort was success-
ful. Despite our PICU already having high compliance
the intervention period. Notably, in the preperiod from with the documented daily discussion of CVC need, we
October 2015 to September 2016, the monthly utilization found that at least 10% of CVCs had potentially mod-
trend had a slope that was not significantly different from ifiable indications. Education around daily discussion of
zero (slope −0.00045, P = 0.59). The median duration of need should continue.
modifiable CVCs was 43 days (interquartile range [IQR] One must interpret the results of our QI initiative in
63) in the first 6 months of the intervention (October light of some limitations. Our single-center experience
2016 to March 2017) compared to a shorter duration of may not be generalizable to other units with a different
32 days (IQR 66) in the second 6 months (April 2017 to structure, size, or patient population. The indications doc-
September 2017). umented for the CVC at the time of placement were at the
provider’s discretion and not always complete. For exam-
ple, there may have been more than one indication (eg,
DISCUSSION PN and difficult vascular access), although only one indi-
The PICU QI team chose to target the meaningful use of cation may have been documented. Also, the modifiable
CVCs, explicitly aiming to optimize the indications and and nonmodifiable groups were a priori determined by
eliminate avoidable continuation of indwelling CVCs in the CVC stewardship team, and improvements in those
the PICU, termed “CVC stewardship.” In particular, we definitions over subsequent cycles have clarified the termi-
targeted the duration of the use of long-term CVC intend- nology. Unlike in-person interviews with providers during
ing to decrease overall long-term CVC days in the PICU. the intervention period, we relied on a chart review for
Following a multifaceted QI initiative that included the preintervention period, which may have less accu-
transparency, increased awareness, and direct in-person rately recorded all the indications for CVC use. The lack
discussions, we demonstrated a significantly decreasing of assessment of balancing measures (eg, whether a new
CVC utilization trend. The project resulted in a culture CVC was required within 24 h of recommended removal)
that promoted proactive discussion related to the need is another limitation. Last, we were unable to collect the
for CVCs, clarifying reasonable indications for continued duration of CVCs for the nonmodifiable CVCs given the
utilization of CVCs in patients, exploring other options, lack of clarity on the timing of removal in the electronic
and considering the replacement of percutaneous CVCs medical record, so we cannot address the total CVC dura-
with PICCs or tunneled CVCs. tion for all central lines, which could help clarify the effect
Continued use of CVCs beyond the absolute min- of our interventions.
imum necessary duration is a risk factor for CLABSI. There were also strengths to our process, including the
However, there are not well-documented efforts to tar- multiple iterations and the ability to include provider
get CVC duration to decrease CLABSI rates in the liter- feedback directly over multiple PDSA cycles. Local lead-
ature, particularly in pediatrics. In a small study of adult ership support and their presence at the front lines of
patients in Thailand, the authors reported a statistically this stewardship intervention were pivotal factors in the
significant decrease in the number of CVC days follow- success of this QI initiative. The variability among our
ing an intervention that promoted physician documen- patient population and provider population, including
tation of the indication for the ongoing use of CVC.15 In 3 distinct medical teams, helped ensure a more gener-
another recent study from the United States in an adult alizable group of patients than might be seen in a more
long-term acute care facility, weekly multidisciplinary focused unit such as a cardiac ICU. The detailed infor-
infection prevention team rounds evaluating CVCs com- mation collected and the prospective nature of CVC
municated recommendations for removal to the primary indications were also beneficial in truly understanding
physician, resulting in nonsignificant improvement in in real-time what was driving the modifiable indications
CLABSI rates over the study timeframe of four months, and how we could enhance education to improve the
without the ability to sustain the change.16 Multiple removal of idle CVCs.

7
 Impact of a Stewardship Intervention on Central Venous Catheter Utilization Pediatric Quality and Safety

CONCLUDING SUMMARY line-associated blood stream infections–United States, 2001, 2008,

and 2009. Ann Emerg Med. 2011;58:447–451.
We have demonstrated a positive influence of our QI 5. Guerin K, Wagner J, Rains K, et al. Reduction in central line-asso-
effort on CVC utilization. Our results emphasize the ciated bloodstream infections by implementation of a postinsertion
care bundle. Am J Infect Control. 2010;38:430–433.
importance of a multidisciplinary QI effort that includes 6. Janum S, Zingg W, Classen V, et al. Bench-to-bedside review: chal-
leveraging electronic health records to promote data- lenges of diagnosis, care and prevention of central catheter-related
driven awareness, practice modifications, and outcome bloodstream infections in children. Crit Care. 2013;17:238.
Downloaded from http://journals.lww.com/pqs by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWn

7. O’Grady NP, Alexander M, Burns LA, et al; Healthcare Infection

improvement. The proactive discussions around the Control Practices Advisory Committee. Guidelines for the pre-
actual need for ongoing use of CVCs may help elimi- vention of intravascular catheter-related infections. Am J Infect
nate some of the barriers to the timely removal of CVCs Control. 2011;39(4 suppl 1):S1–S34.
8. Miller MR, Griswold M, Harris JM II, et al. Decreasing PICU cath-
in the PICU population. Ideally, in the future, we would eter-associated bloodstream infections: NACHRI’s quality transfor-
explore the impact of information technology on reduc-
YQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdgGj2MwlZLeI= on 11/10/2023

mation efforts. Pediatrics. 2010;125:206–213.

ing healthcare-associated infections in busy and complex 9. Giordano P, Saracco P, Grassi M, et al; Italian Association of
Pediatric Hematology and Oncology (AIEOP). Recommendations
healthcare systems. for the use of long-term central venous catheter (CVC) in children
with hemato-oncological disorders: management of CVC-related
occlusion and CVC-related thrombosis. On behalf of the coagu-
DISCLOSURE lation defects working group and the supportive therapy work-
ing group of the Italian Association of Pediatric Hematology and
The authors have no financial interest to declare in relation Oncology (AIEOP). Ann Hematol. 2015;94:1765–1776.
to the content of this article. 10. Wylie MC, Graham DA, Potter-Bynoe G, et al. Risk factors for cen-
tral line-associated bloodstream infection in pediatric intensive care
units. Infect Control Hosp Epidemiol. 2010;31:1049–1056.
11. Burdeu G, Currey J, Pilcher D. Idle central venous catheter-days
ACKNOWLEDGMENTS pose infection risk for patients after discharge from intensive care.
Am J Infect Control. 2014;42:453–455.
This quality improvement project was funded, in part, 12. Leis JA, Shojania KG. A primer on PDSA: executing plan-do-
by the Department of Anesthesiology, Critical Care and study-act cycles in practice, not just in name. BMJ Qual Saf.
Pain Medicine’s Division of Critical Care Medicine funds 2017;26:572–577.
13. Niedner MF, Huskins WC, Colantuoni E, et al. Epidemiology of
for quality improvement. We acknowledge the ongo- central line-associated bloodstream infections in the pediatric inten-
ing commitment to patient safety and quality improve- sive care unit. Infect Control Hosp Epidemiol. 2011;32:1200–1208.
ment of Boston Children’s Hospital and the Medical/ 14. Improving Healthcare with Control Charts: Basic and Advanced
SPC Methods and Case Studies. Carey RG, Staker LV, eds. 1st ed.
Surgical Intensive Care Unit, whose Quality Improvement Milwaukee, Wis.: ASQ Quality Press; 2003.
Committee remains dedicated to improving patient safety. 15. Rattanaumpawan P, Teeratorn N, Thamlikitkul V. A Cluster-
Randomized Controlled Trial of the catheter reminder and evalu-
ation program. Infect Control Hosp Epidemiol. 2016;37:231–233.
16. Chandramohan S, Navalkele B, Mushtaq A, et al. Impact of a mul-
REFERENCES tidisciplinary infection prevention initiative on central line and uri-
1. Goudie A, Dynan L, Brady PW, et al. Attributable cost and length of nary catheter utilization in a long-term acute care hospital. Open
stay for central line-associated bloodstream infections. Pediatrics. Forum Infect Dis. 2018;5:ofy156.
2014;133:e1525–e1532. 17. Arora N, Patel K, Engell CA, et al. The effect of interdisciplinary
2. Marschall J, Mermel LA, Classen D, et al. Strategies to prevent cen- team rounds on urinary catheter and central venous catheter days
tral line-associated bloodstream infections in acute care hospitals. and rates of infection. Am J Med Qual. 2014;29:329–334.
Infect Control Hosp Epidemiol. 2008;29(suppl 1):S22–S30. 18. Ilan R, Doan J, Cload B, et al. Removing nonessential central venous
3. Dudeck MA, Weiner LM, Allen-Bridson K, et al. National Healthcare catheters: evaluation of a quality improvement intervention. Can J
Safety Network (NHSN) report, data summary for 2012, Device- Anaesth. 2012;59:1102–1110.
associated module. Am J Infect Control. 2013;41:1148–1166. 19. Seguin P, Laviolle B, Isslame S, et al. Effectiveness of simple daily
4. Liang SY, Marschall J. Update on emerging infections: news from sensitization of physicians to the duration of central venous and uri-
the Centers for Disease Control and Prevention. Vital signs: central nary tract catheterization. Intensive Care Med. 2010;36:1202–1206.