Keywords
sepsis - emergency medicine - medical order entry systems - quality of health care
Background and Significance
Background and Significance
Sepsis is a life-threatening condition that can lead to organ dysfunction and tissue
hypoperfusion resulting from a dysregulated host response to an infection. It is a
leading cause of morbidity and mortality in the United States and is associated with
roughly 750,000 deaths annually.[1] There is a strong body of evidence demonstrating that early identification and management
of patients with suspected sepsis and septic shock significantly improves patient
outcomes and reduces mortality.[2]
[3]
[4]
[5]
[6] The 2016 Surviving Sepsis Campaign guidelines created two sets of standards known
as the 3- and 6-hour “bundles” ([Table 1]), consisting of elements of care that when implemented as a group, have been shown
to improve patient outcomes.[3] Items that must be completed within 3 hours of patient presentation to a health
care facility (i.e., the “3-hour bundle”) include: (1) measuring a lactate level;
(2) obtaining blood cultures prior to administration of antibiotics; (3) administering
broad spectrum antibiotics; and (4) administering 30 mL/kg of crystalloids for hypotension
or a lactate level of ≥ 4 mmol/L.[3] These guidelines strongly recommend the initiation of antimicrobials within 1 hour
of severe sepsis, as research shows that each hour delay in antibiotic administration
is associated with a 7.6% increase in mortality.[3]
[7]
[8] In fact, in 2018 there was an update to the 2016 sepsis guidelines that recommended
combining the 3-hour bundle and parts of the 6-hour bundle into a combined 1-hour
bundle with the intention to expedite patient resuscitation and management.[9] As of now, the 3-hour and 6-hour sepsis bundles are Centers for Medicare and Medicaid
Services (CMS) core measures that are expected to influence future hospital reimbursement.
Table 1
Surviving sepsis campaign 3-hour and 6-hour bundles[3]
|
To be completed within 3 hours of time to presentation
|
○ Measure lactate level
|
|
○ Obtain blood cultures prior to administration of antibiotics
|
|
○ Administer broad spectrum antibiotics
|
|
○ Administer 30 mL/kg crystalloid for hypotension or lactate ≥ 4 mmol/L
|
|
To be completed within 6 hours of time to presentation
|
○ Apply vasopressors (if needed) to achieve MAP ≥ 65 mm Hg
|
|
○ Reassess fluid status if persistent hypotension
|
|
○ Remeasure lactate if initial level was elevated
|
Abbreviation: MAP, mean arterial pressure.
The emergency department (ED) is a main point of entry for patients with sepsis and
therefore plays an integral role in assisting hospitals to meet the 3-hour bundle
criteria.[10] Historically, health care institutions have struggled to promptly identify and treat
patients with suspected sepsis.[7] Multiple institutions have therefore begun utilizing technology to assist with the
timely surveillance and treatment of patients with severe sepsis.[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18] One study showed that incorporating an electronic order-set for patients with suspected
sepsis into their ED workflow resulted in decreased time to antibiotic administration,
increased percentage of drawing two sets of blood cultures prior to administering
antibiotics, and more appropriate antibiotic selection.[10] In October 2015, the hospital system implemented an optional ED electronic physician
order-set aimed to assist with the early identification and treatment of patients
with suspected sepsis. All ED physicians within the health system were made aware
of this order-set through ED provider meetings, email communications, and word of
mouth. Since its initiation, there has been limited evaluation of the order-set to
determine if it is helping to improve compliance with the 3-hour bundle criteria.
Objective
The goal of this project was to evaluate if the physician order-set improved time
to sepsis treatment in the ED by comparing the time to broad-spectrum antibiotic ordering
in patients whose physicians utilized the electronic order-set to those whose physicians
did not utilize the order-set. It was hypothesized that physician utilization of the
ED sepsis order-set results in a more rapid initiation of broad-spectrum antibiotics
in patients with severe sepsis.
Methods
Population
A retrospective chart review was performed on four affiliated community hospitals
around Pittsburgh, Pennsylvania, United States. All four hospitals had access to the
same electronic physician order-set. Adults at least 18 years old who had International
Classification of Diseases, 10th Edition (ICD-10) diagnosis codes (R65.20; R65.21)
of severe sepsis (defined as sepsis with evidence of organ dysfunction) and/or septic
shock (defined as sepsis with systolic blood pressure < 90 mm Hg or lactate ≥ 4 mmol/L)
from May to July 2016 were included in the project ([Fig. 1]). Patients were excluded if they were not capable of meeting the CMS bundle criteria:
(1) patients not admitted through the ED (e.g., direct admits, transfers, or patients
with scheduled surgery) since there was no opportunity for the electronic physician
order-set to be utilized; (2) if sepsis was due to infectious causes that would not
require CMS-approved broad-spectrum antibiotics for empiric sepsis treatment, such
as Clostridium difficile infection; and (3) if systemic inflammatory response syndrome (SIRS) criteria was
not met during their hospitalization or if it was met greater than 3 hours after ED
admission. The time that patients met SIRS criteria was defined as the time at which
at least two of the four following criteria were met in the ED as documented in the
electronic health record: heart rate (HR) > 90 beats per minute (bpm); respiratory
rate (RR) > 20 bpm; white blood cell > 12,000 or < 4,000 × 103/µL; or temperature > 38.3°C or < 36°C. Time to ED triage was defined as the earliest
documented time of any vital recorded in the electronic health record. This project
was submitted to the Institutional Review Board but was deemed to be a quality improvement
project due to comparison of outcomes for patients within a single institution. Therefore,
it was submitted to and approved by the hospital system's Quality Improvement Review
Committee.
Fig. 1 Patient selection flow diagram.
Outcomes
The primary outcome was the time from positive SIRS criteria to ordering of CMS-approved
broad-spectrum antibiotics ([Appendix A]). Time to ordering antibiotics was chosen instead of time to antibiotic administration
because many extraneous factors, such as diagnostic testing and difficulties obtaining
intravenous access, can potentially delay the administration time of antibiotics once
they are ordered. Secondary outcomes included time from SIRS criteria to antibiotic
administration, time from SIRS criteria to lactate test, hospital length of stay,
and patient disposition upon discharge. Time of SIRS onset was used as time zero for
the outcome intervals because the electronic health record does not document the time
of ED presentation.
Appendix A
List of common CMS-approved broad-spectrum antibiotics for empiric monotherapy of
sepsis[19]
|
Antibiotic class
|
Approved antibiotics for monotherapy
|
|
Penicillins
|
Piperacillin/tazobactam
|
|
Ampicillin/sulbactam
|
|
Cephalosporins
|
Ceftriaxone
|
|
Cefepime
|
|
Ceftaroline
|
|
Carbapenems
|
Meropenem
|
|
Ertapenem
|
|
Fluoroquinolones
|
Levofloxacin
|
Abbreviation: CMS, Centers for Medicare and Medicaid Services.
Data Collection
All data were collected via the inpatient electronic health record. For the primary
and secondary outcomes, data that were retrieved included use of the electronic physician
order-set, time of meeting SIRS criteria, time of ED triage, time of antibiotic ordering,
time of antibiotic administration, time of lactate test, hospital length of stay,
and patient disposition. SIRS criteria were collected to determine time of sepsis
presentation instead of the quick Sepsis-Related Organ Failure Assessment (qSOFA)
score because CMS has not yet adopted qSOFA for its reporting system. Other covariate
data that were collected included patient age, sex, and time and date of ED admission.
Statistical Analysis
Basic descriptive statistics were used to describe the patient characteristics between
the groups either receiving (n = 45) or not receiving (n = 78) the electronic order-set. For continuous variables, means were computed if
data were normally distributed, otherwise medians and interquartile ranges were used
if data were skewed. Either parametric or nonparametric statistical tests were used
to univariately compare patient characteristics between the two groups. Results of
these univariate tests showed no statistical differences at p > 0.05. However, age, SIRS criteria met, and admission time had p-values less than 0.20 and these were still entered into separate multiple regression
models for each time outcome. Standard errors estimated from the multiple regression
models were used to construct the 95% confidence intervals (CIs). If two 95% individual
CIs overlapped, then results would be considered not statistically significant (p > 0.05).
Eight cases were omitted because they did not meet the inclusion criteria for SIRS
being met within 3 hours of ED presentation. The patient characteristics between the
8 excluded and the remaining 123 included patients were compared for differences in
patient characteristics. For all the patient characteristics (except use of the order-set),
the 8 cases were approximately evenly distributed. Regarding use of the order-set,
7 of the 8 cases did not get the order-set and no further analysis was done.
Results
There were 602 patients admitted with an ICD-10 diagnosis code for sepsis in May,
June, and July 2016, and 131 were eligible for chart review after accounting for exclusion
criteria ([Fig. 1]). The overall mean values for time of SIRS criteria to antibiotic order and administration
were imputed for the missing data in one case because a broad-spectrum antibiotic
was never ordered. It was decided to include this individual in the analysis to maintain
the “intention to treat” principle by not subjectively removing any cases. Because
utilization of the order-set was not mandatory, multiple regression analyses were
used to estimate adjusted mean times to ordering. Physicians utilized the order-set
in 45 (36%) of the patients who were admitted through the ED with suspected sepsis
([Table 2]). A majority of the patients included were female (59.5%) and 101 patients (77%)
presented to the ED between 8 am and 12 am. The most common combination of two SIRS criteria met was HR > 90 bpm and RR > 20
bpm. Out of the 131 patient charts reviewed, 45 (34.4%) patients had ED physicians
who utilized the electronic order-set.
Patients whose physicians utilized the order-set were ordered broad-spectrum antibiotics
on average 20 minutes sooner than patients whose physicians did not utilize the order-set
(98.9 minutes, 95% CI: 69.4–128.4 vs. 119.1 minutes, 95% CI: 91.4–146.7); however,
this difference was not found to be statistically significant (p > 0.05). Use of the order-set resulted in reduced mean time to antibiotic administration
(144.5 minutes, 95% CI: 107.5–181.4 vs. 182.4 minutes, 95% CI: 125.8–239.0) and reduced
median time to lactate test (12 minutes, range: 0–20 vs. 18.5 minutes, range: 8–34),
similarly the differences were not statistically significant (p > 0.05). CIs based on the median were used to compare time to lactate test because
roughly 25% of lactate tests were drawn prior to SIRS criteria being met, which would
have resulted in a negative mean time. Patients who had the order-set utilized were
found to have longer hospital lengths of stay (8.4 vs. 7.3 days). There was no difference
(p > 0.05) in patient disposition upon discharge between the two groups. Results of
multivariate analyses did not show statistical significance between the groups for
either the time from SIRS criteria to antibiotic ordering, antibiotic administration,
or time from triage. The means and CIs are all shown in [Table 3].
Table 2
Participant characteristics organized according to use of the electronic order-set
|
Variable
|
Use of electronic order-set
(N = 123)
|
|
Yes (n = 45)
|
No (n = 78)
|
|
Male sex (frequency, %)
|
21 (47)
|
32 (41)
|
|
Age, y (mean, SD)
|
74.8 ± 15.3
|
69.3 ± 15.7
|
|
SIRS criteria met (frequency, %)
|
|
HR, RR
|
21 (47)
|
43 (55)
|
|
HR, WBC
|
6 (13)
|
18 (23)
|
|
HR, Temperature
|
10 (22)
|
9 (12)
|
|
All other combinations
|
8 (18)
|
8 (10)
|
|
ED admission time (frequency, %)
|
|
00:00–08:00
|
11 (25)
|
11 (14)
|
|
08:00–16:00
|
21 (47)
|
28 (36)
|
|
16:00–00:00
|
13 (29)
|
39 (50)
|
Abbreviations: ED, emergency department; HR, heart rate; RR, respiratory rate; SD,
standard deviation; SIRS, systemic inflammatory response syndrome; WBC, white blood
cell.
Note: Data represents frequency and percentage unless otherwise specified.
SIRS criteria: HR (heart rate) > 90 beats per minute (bpm), RR (respiratory rate) > 20
bpm, WBC > 12,000 or < 4,000, temperature > 38.3°C or < 36°C.
Table 3
Outcomes organized according to use of the electronic order-set
|
Outcomes
|
Use of electronic order-set (N = 123)
|
|
Yes (N = 45)
|
No (N = 78)
|
|
Mean
|
95% CI
|
Mean
|
95% CI
|
|
SIRS to antibiotic order, minutes[a]
|
98.9
|
(69.4–128.4)
|
119.1
|
(91.4–146.7)
|
|
SIRS to antibiotic administration, minutes[a]
|
144.5
|
(107.5–181.4)
|
182.4
|
(125.8–239.0)
|
|
Triage to antibiotic order, minutes[a]
|
113.7
|
(80.4–146.9)
|
145.9
|
(120.9–170.8)
|
|
SIRS to lactate level, min[b] (median, 95% CI for median)
|
12
|
0–20
|
18.5
|
8–34
|
|
Length of stay, days[a]
|
8.4
|
(5.7–11.0)
|
7.3
|
(6.1–8.4)
|
|
Disposition (frequency, %)
|
|
Home
|
11 (24)
|
26 (33)
|
|
Skilled nursing facility or rehabilitation
|
22 (49)
|
34 (44)
|
|
Death
|
12 (27)
|
18 (23)
|
Abbreviations: CI, confidence interval; SIRS, systemic inflammatory response syndrome.
a Mean values and their confidence intervals represent the least squares estimates
from the multiple regression models, adjusted for age, SIRS criteria, and admission
time.
b Data represent the median values displayed. Exact statistical methods were used to
obtain the confidence intervals.
Discussion
In this multicenter retrospective chart review, we found that the use of the ED electronic
order-set did not show a significant reduction in time in either antibiotic ordering
(∼20 minutes) or administration (∼38 minutes). Patients treated with the order-set
had a mean time to antibiotic administration of 144.5 minutes (95% CI: 107.5–181.4)
compared with 182.4 minutes (95% CI: 125.8–239.0) without the order-set. Regardless
of the lack of statistical significance, the electronic order-set did reduce the time
to antibiotic administration in patients with suspected sepsis, and the results suggest
that the utilization of this order-set may help hospitals meet the CMS core measure
of administering antibiotics within 3 hours (180 minutes) of sepsis presentation,
but further research is required.
Strengths of this project include it being multicenter and having objective outcome
measures. Limitations include it being retrospective, having a high risk of selection
bias due to the order-set being optional, low physician utilization of the order-set,
and the lack of generalizability to other hospitals since the order-set is specific
to only one health system. Another major limitation was the small sample size, which
could have influenced the results showing nonstatistically significant changes. There
were also numerous confounders that could have potentially affected the results, such
as patient acuity, voluntary use of the order-set by physicians, and physician comfort
level with using electronic order-sets. Patients who had the order-set utilized on
average had a 1-day longer length of stay, suggesting that physicians may have preferentially
chosen to utilize the order-set in higher acuity patients. Another potential confounder
is the design of the order-set. Generally, assessment and management of sepsis is
divided into two phases: (1) patient work-up (checking vitals, ordering laboratories
and cultures, etc.) and (2) treatment (fluid resuscitation, antibiotic selection,
etc.). This specific order-set is monophasic, in which physicians are prompted to
order all the laboratories required for patient assessment at the same time as being
prompted to order antibiotics and fluids. As a result, it is possible that physicians
are initially utilizing the order-set for the recommended laboratories and cultures,
but may not be returning to the order-set once it is time to order antibiotics. This
could result in extended time to antibiotic ordering.
Overall, this review demonstrates that the use of electronic order-sets in the ED
has the potential to reduce both time to antibiotic ordering and administration in
patients with suspected sepsis, and potentially improve hospitals' compliance with
meeting the Sepsis CMS Core Measures.
Conclusion
The use of the electronic order-set was associated with reduced time to antibiotic
ordering and administration in patients with suspected sepsis in the ED; however,
the findings were not statistically significant.
Clinical Relevance Statement
Clinical Relevance Statement
This project looked at the efficacy of an electronic tool that may be used to help
improve patient outcomes in addition to helping the hospital system meet the Sepsis
CMS Core Measure. Even though this specific order-set did not show statistically improved
time to antibiotic ordering and administration in patients with severe sepsis, electronic
order-sets in the ED have great potential to help hospital systems meet the Sepsis
CMS Core criteria 3-hour bundle. Low utilization of the sepsis order-set by physicians
may require further investigation.
Multiple Choice Questions
Multiple Choice Questions
-
When implementing an electronic physician order-set in the emergency department, which
of the following is more important to take into consideration when designing the order-set?
-
Physician input.
-
Nursing input.
-
eRecord input.
-
Administration input.
Correct Answer: The correct answer is option a, physician input. Since the physician will be the
only provider utilizing the order-set, it is important that they provide input to
assist in the design of the order-set so that it matches their workflow. This may
also encourage more physicians to utilize the order-set since they would assist with
the design.
-
Which of the following is a criterion for the Surviving Sepsis Campaign 3-hour bundle?
-
Measure procalcitonin level.
-
Administer broad spectrum antibiotics.
-
Obtain blood cultures after administering antibiotics.
-
Administer albumin replacement for hypotension and resuscitation.
Correct Answer: The correct answer is option b, Administer broad spectrum antibiotics. Administration
of broad spectrum antibiotics is one of the 3-hour bundle criteria, that is now being
used by CMS as a quality member that influences hospital reimbursement for patients
admitted with sepsis. The other criteria of the 3-hour bundle are: (1) measure lactate
level; (2) obtaining blood cultures prior to administration of antibiotics; and (3)
administer 30 mL/kg of crystalloid solution (albumin is a colloid) for hypotension
or lactate ≥ 4 mmol/L.
-
What is the most important reason for developing an electronic order-set to be used
in the emergency department to aid in the management of patients with sepsis?
Correct Answer: The correct answer is option d, Improve patient safety and outcomes. Everything in
health care should be done with the primary focus of improving patient care. Reducing
health care spending and optimizing hospital reimbursement is very important, but
the patient should always be the first priority.
