Article, Emergency Medicine

Emergency department sepsis screening tool decreases time to antibiotics in patients with sepsis

a b s t r a c t

Recent literature has highlighted the importance of early identification and treatment of sepsis; however, limited data exists to help recognize sepsis in the emergency department (ED) through use of a screening tool. The pur- pose of this study was to evaluate the impact of a sepsis screening tool implemented in an academic medical cen- ter ED on compliance with the 3-hour Sepsis Bundle.

This was a retrospective cohort study that included a total of 115 patients, of which 58 were in the pre-tool group and 57 were in the post-tool group. There was no difference in 3-hour bundle compliance between groups (36.2% vs. 47.4%, P = 0.26). There was no difference in the following bundle components: lactate (79.3% vs. 80.7%, P = 0.85), blood cultures (86.2% vs. 96.5%, P = 0.09), blood cultures before administering antibiotics (91.4% vs. 100%, P = 0.57) and adequate fluids administration (44.7% vs. 41.9%, P = 0.820). A significantly higher number of pa- tients received antibiotics within 3 h in the post-tool group (58.6% vs. 89.5%, P b 0.001). Statistically significant secondary outcomes included average time to antibiotics (P = 0.04), administering antibiotics within an hour (P N 0.001), and ICU length of stay (P = 0.03). There was no difference in 30-day mortality, however mortality was numerically lower in the post-tool group (36.2% vs. 26.3%, P = 0.25).

Although implementation of an ED sepsis screening tool did not increase 3-hour bundle compliance, it did in- crease the proportion of patients receiving timely antimicrobial therapy and demonstrated a trend towards de- creased mortality.

(C) 2018


Sepsis is a dysregulated host response to infection which may lead to organ dysfunction. Septic shock occurs with progression to organ dys- function and is associated with a high risk of mortality [1,2]. Sepsis is a common cause for hospitalization with N750,000 cases annually, with majority of cases presenting through the emergency department (ED) [3,4]. In order to increase recognition and decrease sepsis-related mor- tality, the Surviving Sepsis Campaign (SSC) guidelines specify care bun- dles, at 3 and 6 h after recognition, targeting early management of severe sepsis and septic shock. The 3-hour bundle includes measuring a Serum lactate level, obtaining blood cultures prior to antibiotics ad- ministration and administering broad spectrum antibiotics in addition to a bolus of 30 mL/kg of crystalloids for hypotension or a serum lac- tate >= 4 mmol/L. [5] The guidelines recommend administering antibi- otics within the first hour of recognizing severe sepsis and septic shock [5].

* Corresponding author at: Loyola University Medical Center, 2160 S. 1st Ave, Maywood, IL 60153, USA.

E-mail address: [email protected] (T. Shah).

Early recognition and compliance with the SSC bundle has demon- strated a Mortality benefit [1,5-7]. While some Intensive care units and Inpatient floors have implemented a sepsis screening tool to help health care providers identify septic patients, there is limited data on the use of Screening tools in the ED setting. A commonly used inpatient screening tool, the modified early warning score , is based on five physiologic parameters to detect clinical deterioration [8]. MEWS helps identify patients who require increased level of care early, which can increase compliance with the 3-hour bundle. Some ICUs have developed unit-specific sepsis screening tools. One retrospec- tive study in a surgical ICU (SICU) observed a decrease in mortality rate from 35.1% to 23.3% in 136 patients after implementation of a sepsis screening tool [9].

While several screening tools have been evaluated in an inpatient setting, [7-9] few studies have explored the impact of a sepsis screening tool in the ED. One retrospective study assessed the impact of a sepsis Quality improvement project on compliance with the SSC bundle in the ED [10]. The project utilized a computer-assisted screening algo- rithm that generated a pop-up alert to healthcare providers along with a sepsis order set to assist with initial management. The study found a statistically significant improvement in time to antibiotics, time to intravenous fluids and 3-hour bundle compliance; however,

0735-6757/(C) 2018

1746 T. Shah et al. / American Journal of Emergency Medicine 36 (2018) 17451748

there was no difference in mortality. Currently, there are no widely- used, validated ED sepsis screening tools. The purpose of this study was to evaluate the impact of the sepsis screening tool implemented in our ED to determine its impact on 3-hour bundle compliance in pa- tients with sepsis.


This retrospective cohort study conducted at Loyola University Med- ical Center, an academic medical center with approximately 23,800 ad- missions and N48,000 ED visits annually. ED patients >= 18 years of age, admitted to either the general floor, medical ICU (MICU) or SICU and di- agnosed with severe sepsis and septic shock were included in this study. The pre-tool period was from August 2012 to January 2013 and post- tool period was from January 2015 to June 2015. Patients admitted from 2013 and 2014 were excluded as the screening tool was being im- plemented and modified to its current form this time. The following pa- tients were excluded: trauma, admitted to units other than the general floor, MICU and SICU, without sepsis at time of presentation to the ED, and pregnant women. International Classification of Diseases (ICD) -9 diagnoses codes were used to identify patients and manual chart review was conducted within the electronic medical records to validate accu- racy of coded data and collect baseline variables and bundle compliance data.

The sepsis screening tool was first implemented in May 2013 and updated and automated in December 2014. The screening tool com- prised of two parts: five vital signs and three dichotomous questions. The vital signs were automatically populated each time they were en- tered by the nurse. Recorded measures included: temperature, systolic blood pressure (SBP) or mean arterial pressure, respiratory rate, oxygen saturation, and pulse. The dichotomous questions answered by the nurse included the following: rigors present, suspected infection and al- tered mental status present. If three or more values out of the eight col- lected were abnormal or present, the patient was considered positive for the sepsis screen. The ED had an electronic track board which could be viewed by all health care providers. It displayed all the patients in the ED and waiting room along with Acuity level, age, sex, room num- ber, and chief complaint. The acuity level is presented in a circular icon that changes to a square upon screening positive for sepsis to alert pro- viders. The physician then evaluated the patient and initiated the sepsis order set if sepsis is suspected. The sepsis order set was available during the entire study period, and provided physicians the ability to quickly order appropriate laboratory testing, blood cultures, diagnostic proce- dures, intravenous fluids, and medications.

The primary outcome of this study was the percentage of patients who had the 3-hour bundle completed within 3 h. Secondary outcomes included the individual bundle endpoints (time to lactate, blood cul- tures, broad-spectrum antibiotics, and adequate infusion of fluids [30 mL/kg] if required for hypotension or a lactate >= 4 mmol/L), as well as number of patients who received antibiotics within an hour, time to va- sopressors, number of vasopressors, hydrocortisone use, length of stay (LOS) in the ED, ICU and hospital, and 30-day mortality. Data collection included the following: age, sex, race, comorbidities, Sequential Organ Failure Assessment score, admitting unit, ED shift, systemic in- flammatory response syndrome (SIRS) criteria, severity of sepsis, 3- hour bundle components, screening tool components, and secondary outcomes. Sepsis time zero for all patients was when the triage nurse obtained the first set of vitals.

Data were analyzed using SPSS version 23 (Chicago, IL). Baseline var-

iables displayed using descriptive statistics, including medians, inter- quartile ranges (IQR) and percentages. Continuous variables were analyzed using a t-test if parametric or Mann-Whitney U test if non- parametric. Chi-square and Fisher’s exact test were used for categorical data as appropriate. A multivariate logistic regression analysis was con- ducted using variables with a p-value b 0.2 to determine independent predictors of 30-day mortality. A post-hoc analysis was conducted to

evaluate the new proposed sepsis identification criteria, quick sequen- tial organ failure assessment (q-SOFA), which is comprised of three var- iables: RR, altered mental status and SBP [1]. Results were considered statistically significant if the p-value b 0.05.


A total of 293 patients were reviewed, of which 115 patients (39.2%) met inclusion criteria, (n = 58 pre-tool group, n = 57 post-tool group; Fig. 1). Baseline characteristics were similar between the two groups in- cluding admission time of day, baseline serum lactate, diagnosis and ad- mitting unit (Table 1). The most common comorbidities in both groups were hypertension and diabetes. Source of infection was similar be- tween the two groups, however genitourinary infections were common in the post-tool group (18.9% vs. 36.8% patients, P = 0.028).

There was no difference in the primary outcome of 3-hour bundle compliance between the pre-tool and post-tool groups (36.2% vs. 47.4%; P = 0.225) (Table 2). For the secondary endpoints of individual bundle components, the only significant difference was observed in the proportion of patients receiving broad spectrum antibiotics within 3 h (58.6% vs. 89.5%; P b 0.001). Receipt of antibiotics at 1h after sepsis recognition was also statistically significant between groups (10.3% pre-tool vs. 89.5% post-tool; P b 0.001). The median time to antibiotics was also significantly faster in the post-tool group (144 [IQR 96-234] min vs. 60 [IQR 30-96] min, P b 0.001).

There were no differences in other secondary endpoints between groups, including vasopressor use within 6 h, hydrocortisone adminis- tration, appropriate selection of antibiotics according to suspected source of infection, ED LOS, hospital LOS or 30-day mortality (Table 3). ICU LOS was significantly shorter in the post-tool group (4 [IQR 1.5-14] days vs. 3 [IQR 0-6] days, P = 0.027).

A multivariate analysis was performed to determine independent predictors of 30-day mortality. Variables included from the univariate analysis were blood cultures within 3 h, elevated lactate levels, ade- quate fluid resuscitation as determined by the SSC guidelines (if hypo- tensive or serum lactate >= 4 mmol/L) and sepsis screening tool group. The analysis demonstrated that initial lactate level N 4 mmol/L signifi- cantly impacts 30-day mortality, (95% CI, 1.52 to 7.92; P = 0.003), how- ever the screening tool did not appear to affect mortality.


Our study evaluated the impact of an ED sepsis screening tool. While an increase in compliance with the SSC 3-hour bundle was not ob- served, there was a numerical increase in bundle compliance. Impor- tantly, we observed an increase in antibiotic administration within 1 h and 3 h by 30.9% and 79.2%, respectively, in the post-tool group. This has important implications as delayed antimicrobial therapy following documented hypotension has been associated with increased mortality. One study demonstrated administration of antibiotics within the first hour of septic shock was associated with a 79.7% survival to hospital dis- charge as opposed to 60% survival rate if antibiotics were administered within 3 h [6]. Delayed antibiotic therapy has also been associated with increased LOS, acute kidney injury, acute lung injury and increased se- verity of illness as determined by SOFA score [6,11-15].

This study evaluated the impact of an ED sepsis screening tool on adult patients. The tool was created using triage Nurse assessment and electronic cues to the ED clinicians in an effort to expedite recognition and treatment of sepsis. A similar retrospective cohort study evaluated a sepsis screening tool (n = 624), and found an increase in bundle com- pliance post-implementation (p b 0.001), though no difference was ob- served in mortality. This study utilized SIRS criteria in the screening tool, which is non-specific and less accurate than SOFA-based assessments. Our assessment tool utilized a combination of vital signs and screening questions, which allowed for faster assessment as it was not reliant on laboratory parameters (e.g. white blood count) [10]. This previous

T. Shah et al. / American Journal of Emergency Medicine 36 (2018) 17451748 1747

Fig. 1. Study flow chart.

study also implemented a sepsis order set concomitantly with their screening tool, which may have contributed to the improvement in bundle compliance observed.

A quasi-experimental study of patients with septic shock admitted to an academic tertiary center analyzed the impact of a hospital-wide quality improvement program created based on the SSC bundle com- pared to a historical control group [7]. This program included education via conferences, lectures, posters, pocket cards and emphasized early recognition and interventions. A total of a 384 patients were included. In-hospital mortality was significantly reduced in the intervention group (57.3% vs. 37.5%, P = 0.001) and remained significant after ac- counting for confounding factors. This difference remained significant after controlling for confounding factors (odds ratio, 0.50; 95% confi- dence interval, 0.28-0.89). This differs from our patient population as subjects were inpatient and only septic shock patients were included.

Table 1

Baseline characteristics.

Pre-tool (n = 58)

Post-tool (n = 57)


Age, years, median (IQR)

61.5 [IQR 55-75]

67 [IQR 55-80]


Male, n (%)

34 (58.6)

32 (56.1)


Race, n (%)


32 (55.2)

29 (50.9)


African American

18 (31.0)

14 (24.6)



8 (13.8)

12 (21.1)



Comorbidities (N 25%), n (%)


1 (1.8)



20 (37.0)

20 (35.1)



27 (50.0)

32 (56.1)



8 (14.8)

16 (28.1)


Kidney dysfunction

16 (29.6)

12 (21.1)



13 (24.1)

14 (24.6)


Cardiovascular disease

9 (16.7)

18 (31.6)


This study supports initiatives to promote early recognition and treat- ment of sepsis.

Our study has several limitations including its single-centered, retro- spective design. Data collection was reliant on appropriate nursing docu- mentation. For example, in the post-tool group, only 91.2% patients had a positive sepsis screening score. This was due to an error in answering the tool questions or a quick change in the patient’s clinical status. All health care providers continue to receive ongoing education on using the sepsis screening tool and order set. Our screening tool was implemented in order to facilitate rapid detection of patients who may be septic, but its accuracy has not been validated. Furthermore, variables were selected prior to the creation of quick SOFA (qSOFA) assessment as described in the new sepsis-3 guidelines [1]. We plan to evaluate our tool against qSOFA in future studies. This study’s pre/post implementation design may have led to a chronological bias due to an increase in sepsis aware- ness and focus in light of recent landmark trials [4,16,17]. Another limita- tion is the inability to quantify ED patient volume and variation in the speed of the process of ordering to administering medications. Finally, we had a relatively small sample size, which may have hindered our abil- ity to detect a difference in bundle compliance between groups.


Sepsis screening tools are not a universal practice in the ED, but may help to identify severe septic and septic shock patients early in order to initiate treatment and meet the 3-hour bundle components. Our study demonstrated that while overall 3-hour bundle compliance was not im- proved, appropriate antimicrobials were initiated faster. Further studies are required to evaluate the implementation of screening tools in EDs based on the new Sepsis-3 definitions.


SOFA score, median (IQR)

6 [IQR 3-9]

5 [IQR 3-9]


Table 2

ED PM shift (1900-0700 h), n (%)

22 (37.9)

17 (29.8)


3-hour bundle compliance.

Weekend admit, n (%)

13 (22.4)

17 (29.8)


Sepsis tool score >= 3 points

52 (91.2)


(n = 58) (n = 57)

Baseline Lactate, mean +- SD

4.45 +- 4.21

3.69 +- 2.96




SD – standard deviation; SOFA – sequential organ failure assessment; ED – emergency de- partment; MICU – medical intensive care unit; SICU – surgical intensive care unit;

Diagnosis, n (%)


Lactate within 3 h, no. (%)

46 (79.3)

46 (80.7)


Severe sepsis

27 (46.6)

28 (49.1)

Blood cultures within 3 h, no. (%)

50 (86.2)

55 (96.5)


Septic shock

31 (53.4)

29 (50.9)

Blood cultures before antibiotics, no. (%)

53 (91.4)

57 (100)


Admitting unit, no. (%)


Antibiotics within 3 h, no. (%)

34 (58.6)

51 (89.5)



45 (77.6)

43 (75.4)


Received adequate fluids (30 mL/kg) of no.






1 (1.8)


needing fluid resuscitation, no. (%)

General floor

13 (22.4)

13 (22.8)

3-hour bundle compliance, no. (%)

21 (36.2)

27 (47.4)


3-hour bundle compliance excluding fluids, no. (%)

31 (53.4) 43 (75.4) 0.014

1748 T. Shah et al. / American Journal of Emergency Medicine 36 (2018) 17451748

Table 3

Secondary outcomes.

Pre-tool (n = 58)

Post-tool (n = 57)


Wang HE, Shapiro NI, Angus DC, Yealy DM. National estimates of severe sepsis in United States emergency departments. Crit Care Med 2007;35(8):1928-36.
  • Yealy DM, Kellum JA, Huang DT, Barnato AE, Weissfeld LA, Pike F, et al. A randomized trial of protocol-based care for early septic shock. N Engl J Med 2014;370(18): 1683-93.
  • Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. Surviving
  • Vasopressors use in 6 h, no. (%) 26 (44.8) 28 (49.1) 0.644

    Hydrocortisone use, no. (%) 12 (20.7) 17 (29.8) 0.259

    Appropriate antibiotic use, no. (%) 53 (91.4) 54 (94.7) 0.717

    sepsis campaign: international guidelines for management of severe sepsis and sep- tic shock: 2012. Crit Care Med 2013;41(2):580-637.

    Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Duration of hypo-

    Average time to antibiotics, (mins) median (IQR)

    144 [IQR


    60 [IQR



    tension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med 2006;34(6):1589-96.

    Antibiotic administered within an hour, 6 (10.3) 51 (89.5) b0.001 no. (%)

    Sepsis order set use, no. (%) 26 (45.6) 50 (87.7) b0.001

    Castellanos-Ortega A, Suberviola B, Garcia-Astudillo LA, Holanda MS, Ortiz F, Llorca J,

    et al. Impact of the surviving sepsis campaign protocols on hospital length of stay and mortality in septic shock patients: results of a three-year follow-up quasi-

    ED length of stay, (hours), median (IQR) 10.1 [IQR


    6.9 [IQR



    experimental study. Crit Care Med 2010;38(4):1036-43.

    Subbe CP, Kruger M, Rutherford P, Gemmel L. Validation of a modified early warning

    ICU length of stay, (days), median (IQR) 4 [IQR 1.5-14] 3 [IQR 0-6] 0.027

    score in medical admissions. QJM 2001;94(10):521-6.

    Hospital length of stay, (days), median (IQR)

    8 [IQR 4-16] 6 [IQR 4-14] 0.265

    Moore LJ, Jones SL, Kreiner LA, McKinley B, Sucher JF, Todd SR, et al. Validation of a screening tool for the early identification of sepsis. J Trauma 2009;66(6):1539-46 [discussion 46-7].

    30 day mortality, no. (%) 21 (36.2) 15 (26.3) 0.253

    ED – emergency department; ICU -intensive care unit;

    Conflicts of interest

    T. Shah, M.A. Rech, E. Sterk: no conflicts to disclose.

    Author contributions

    T.S. designed the trial, collected, managed and analyzed the data, and drafted the manuscript. E.S. supervised the study design and contrib- uted to the manuscript revisions. M.R. supervised the study design, pro- vided statistical advice and contributed to the manuscript revisions.


    1. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315(8):801-10.
    2. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Crit Care Med 2017;45(3):486-552.
    3. Gatewood MO, Wemple M, Greco S, Kritek PA, Durvasula R. A quality improvement project to improve early sepsis care in the emergency department. BMJ Qual Saf 2015;24(12):787-95.
    4. Ferrer R, Martin-Loeches I, Phillips G, Osborn TM, Townsend S, Dellinger RP, et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement pro- gram. Crit Care Med 2014;42(8):1749-55.
    5. Zhang D, Micek ST, Kollef MH. Time to appropriate antibiotic therapy is an indepen- dent determinant of postinfection ICU and hospital lengths of stay in patients with sepsis. Crit Care Med 2015;43(10):2133-40.
    6. Bagshaw SM, Lapinsky S, Dial S, Arabi Y, Dodek P, Wood G, et al. Acute kidney injury in septic shock: clinical outcomes and impact of duration of hypotension prior to ini- tiation of antimicrobial therapy. Intensive Care Med 2009;35(5):871-81.
    7. Iscimen R, Cartin-Ceba R, Yilmaz M, Khan H, Hubmayr RD, Afessa B, et al. Risk factors for the development of acute lung injury in patients with septic shock: an observa- tional cohort study. Crit Care Med 2008;36(5):1518-22.
    8. Garnacho-Montero J, Aldabo-Pallas T, Garnacho-Montero C, Cayuela A, Jimenez R, Barroso S, et al. Timing of adequate antibiotic therapy is a greater determinant of outcome than are TNF and IL-10 polymorphisms in patients with sepsis. Crit Care 2006;10(4):R111.
    9. Mouncey PR, Osborn TM, Power GS, Harrison DA, Sadique MZ, Grieve RD, et al. Trial of early, goal-directed resuscitation for septic shock. N Engl J Med 2015;372(14): 1301-11.
    10. Peake SL, Delaney A, Bailey M, Bellomo R, Cameron PA, Cooper DJ, et al. Goal- directed resuscitation for patients with early septic shock. N Engl J Med 2014;371 (16):1496-506.

    Leave a Reply

    Your email address will not be published. Required fields are marked *