Secondary Logo

Journal Logo

Single-Center Quality Improvement Report

Bedside Evaluation for Early Sepsis Intervention: Addition of a Sepsis Response Team Leads to Improvement in Sepsis Bundle Compliance

Suliman, Sally MD1; Price, John DO1; Cahill, Meredith MPH2; Young, Taylor MPH2; Furmanek, Stephen MPH2; Galvis, Juan MD1; Shoff, Hugh MD3; Parra, Frankie BSN3; Stevenson, Gina BSN3; Cavallazzi, Rodrigo MD1

Author Information
doi: 10.1097/CCE.0000000000000312


Sepsis is a systemic and detrimental host response to an infection, which often leads to multiple organ dysfunction and shock. It is a major public health concern, affecting more than 1.5 million Americans each year and killing one in four patients. Despite advances in the early recognition and intervention of sepsis, the prevalence continues to be on the rise and sepsis carries an annual cost of $20 billion (1).

Since the study by Rivers et al (2), there have been multiple landmark trials comparing early goal-directed therapy (EGDT) to usual care (3–5). While not all elements of EGDT are beneficial, there are key aspects of patient care in sepsis that have an impact on outcomes and are contemplated in guidelines. The Surviving Sepsis Campaign recommends the use of sepsis bundles to facilitate the implementation of these guidelines. The emphasis is compliance with a seven-element bundle that consists of blood cultures being drawn, an initial serum lactic acid and the administration of antibiotics within a set time frame. It also includes aggressive fluid therapy, repeat assessment of fluid responsiveness, and, if necessary, the use of vasopressors for hemodynamic support.

The early implementation of sepsis bundles can only be achieved if patients with sepsis are recognized. In fact, the International Clinical Practice and Guidelines and the Centers for Medicare and Medicaid Services recommend the prompt identification and treatment of sepsis (6). Early recognition of the septic patient combined with prompt treatment improves morbidity and mortality and could reduce the number of avoidable deaths (7).

There was significant variation in patient outcomes within our institution. This was likely in part due to delayed patient recognition, extended stay in the emergency department (ED), and a lack of standardization to the approach of the septic patient. Prior studies have demonstrated that a multidisciplinary approach increases sepsis bundle compliance, leading to a relative risk reduction in mortality (8,9). Therefore, we decided to implement a multifaceted approach to increase the compliance with the sepsis bundle. This included education of the house staff, an alert system with communication between the ED providers and the ICU house staff, and a dedicated sepsis rapid response system.


Study Design

The study was conducted at the University of Louisville, which is a 404-bed tertiary academic medical center between February 2018 and was completed August 2019. The study included septic patients identified in the ED who were followed to either hospital discharge or death. The proposal was approved by our institutional review board (number 18.0135) as a quality improvement (QI) project, and patient information and data remained protected by the HIPAA.


A prior retrospective chart review was carried out over a 12-month period before the initiation of our project on all patients who were admitted through the ED with a diagnosis of sepsis or septic shock as documented by International Classification of Diseases codes. Compliance with the elements of the sepsis bundle was recorded and revealed an overall compliance rate of only 37%.

At the project, start key committee members were identified and a multidisciplinary QI team was created. The first task was to define the phase in order and create a workflow plan. The team divided the project into three primary areas of focus: 1) the early identification of a septic patient, 2) standardization of early resuscitation including administration of broad-spectrum antibiotics and adequate fluid resuscitation, and 3) appropriate triage of the patient. Individual and overall bundle compliance were to be followed, and feedback provided at monthly meetings to educate and improve physician adherence to the sepsis bundle.

A systemic inflammatory response syndrome (SIRS) trigger and sepsis crawler algorithm were created by our institutions' informatics department to allow for the early identification of a potentially septic patient. These alerts would trigger a page to the ED nurse when a patient met SIRS criteria. The nurse was then responsible for notifying the ED physician who would perform a bedside evaluation and determine if the patient had sepsis or septic shock. Sepsis was defined as those patients meeting two SIRS criteria and an identified source of infection. The ED team was then responsible for obtaining the initial set of blood cultures (and or urine and a chest radiograph depending on the presumed source of infection) before antibiotic administration, as well as an initial lactic acid measurement and then starting necessary antibiotics. Weight-based IV fluid resuscitation was also initiated for all appropriate patients. For each septic patient, the ED simultaneously initiated a call to the sepsis response team (SRT). This SRT consisted of either an ICU resident or fellow who would respond to the patient’s bedside within a 20-minute window and ensure the completion of the 3-hour bundle, initiate steps to ensure a follow-up physical examination and repeat lactic acid were scheduled, as well as appropriately triage the patient. A sepsis checklist of all the interventions with time stamps was also created as part of the project. It was initiated by the ED and continued by the SRT once the patient was triaged. The data was then collected and analyzed biweekly. Results of the preliminary data were given to the SRT as well as the ED to encourage adherence to the bundle and continue to highlight areas for improvement.

Prior to the initiation of the project, pocket cards were made and distributed to the ED nurses, clinicians, as well as the SRT residents, fellows, and attending physicians. These cards highlighted each group’ specific role and criteria for which the SRT should be called. Extensive education of the ED nurses and triage team was also carried out via posters and flow sheets, and ED and SRT residents also received educational material and underwent training sessions before the start of the project (Fig. 1).

Figure 1.:
Sample flow sheet used to educate emergency department staff and nursing. ABX = antibiotics, CXR = chest x-ray, MAP = mean arterial pressure.

Last, we met with the hospital's quality and metrics committee to ensure all necessary quality measures were covered in the checklist and that the workflow was manageable for all groups involved, as well as with the informatics department to set up the sepsis alerts that would trigger the workflow.

Inclusion Criteria

Adult patients presenting with two SIRS criteria plus a suspected source of infection and lactate acid greater than 2 mmol/L or those with two SIRS criteria or hypotension defined as a systolic blood pressure less than 100 mm Hg would trigger a call to the SRT and receive a bedside evaluation. Patients that met criteria for sepsis were included in the study. Patients who had an alternative diagnosis causing SIRS criteria, such as trauma or seizures, or patients transferred from an outlying facility were excluded from the study.


The primary metric outcome was sepsis bundle adherence. We hypothesized this QI initiative would improve adherence with the sepsis bundle elements by 40% in an 18-month period. These elements were 1) initial lactic acid measurement; 2) remeasurement if elevated greater than 2 mmol/L; 3) obtaining blood cultures before the initiation of antibiotics; 4) antibiotic administration within 3 hours of triage; 5) administration of a fluid bolus of 30 mL/kg; 6) evaluation of fluid status using central venous pressure, cardiac ultrasound, or complete physical examination (including pulses and capillary refill time) or central venous oxygen saturation; and 7) if needed vasopressors were started for hemodynamic support. Secondary outcomes included all-cause inhospital mortality, hospital length of stay, and time in the ED.


Pearson chi-square tests were used to assess any differences in the proportion of sepsis bundle compliance between the study population and historical measures, as well as clinical outcomes between patients who had full bundle compliance and those who did not. Differences in patient time spent in the ED and overall length of stay were analyzed by fitting Cox proportional hazards regression models and important predictors of sepsis plan compliance were identified through linear regression. Additionally, two-sided t tests were applied to measure any incongruity in demographics and comorbid conditions among those who received and did not receive the complete sepsis bundle. Analysis was completed using RStudio Version 3.6.1 (2019-07-05). Copyright (C) 2019 The R Foundation for Statistical Computing Platform: x86_64-apple-darwin15.6.0 (64-bit).


The study included a total of 163 patients. Of those, 93 (57%) were male. The mean age was 54 years. There was no significant difference in demographics or comorbid conditions between patients who received all the components of the sepsis bundles and those who did not (Table 1). Intake measurements were also similar between these two groups. Compliance was significantly better if the patient was admitted or later transferred to the ICU (p = 0.039). The most common comorbid conditions were diabetes and diagnosis of cancer.

Table 1. - Baseline Patient Characteristics
Patient Demographics Sepsis Bundle Noncompliance (n = 35) Sepsis Bundle Compliance (n = 128) p
Female, n (%) 18 (51) 52 (41) 0.341
Age, median (IQR) 58 (46–65) 54 (39–66) 0.446
Race/ethnicity, n (%) 0.730
 White 26 (74) 84 (66)
 Black 9 (26) 42 (33)
 American Indian or Alaska Native 0 (0) 1 (1)
 Hispanic/Latino 0 (0) 1 (1)
Intake measurements
 Temperature (°C), median (IQR) 38 (37–39) 37 (37–39) 0.082
 Respiratory rate, median (IQR) 20 (18–24) 22 (18–24) 0.471
 Heart rate, median (IQR) 117 (100–127) 118 (102–133) 0.365
 WBC count, median (IQR) 15 (9–19) 14 (9–18) 0.826
 Initial lactic acid level ≥ 2 and < 4, mmol/L, n (%) 21 (60) 66 (52) 0.344
 Initial lactic acid level ≥ 4 (%), mmol/L, n (%) 5 (14) 34 (27) 0.090
 Systolic blood pressure < 100 mm Hg, n (%) 8 (23) 24 (19) 0.763
 Patient admitted or transferred to the ICU, n (%) 10 (29) 64 (50) 0.039
Comorbidities, n (%)
 Diabetes 10 (29) 35 (27) > 0.999
 End-stage renal disease 0 (0) 4 (3) 0.658
 Congestive heart failure 4 (11) 10 (8) 0.737
 Cerebrovascular accident 2 (6) 9 (7) > 0.999
 IV drug abuse 7 (20) 17 (13) 0.469
 Chronic obstructive pulmonary disease 6 (17) 17 (13) 0.758
 Cancer 10 (29) 28 (22) 0.545
 Cirrhosis 2 (6) 4 (3) 0.83
 HIV 1 (3) 2 (2) > 0.999
 Immunosuppressed 7 (20) 12 (9) 0.150
 Coronary artery disease 4 (11) 15 (12) > 0.999
IQR = interquartile range.

Overall compliance with the sepsis bundle was 79%. There was significant improvement in comparison with the hospital’s historical compliance of 37% (p < 0.001). Compliance with the individual bundle components ranged from 80% to 100%. The components that had lowest compliance were the 30 cc/kg IV fluid bolus followed by documentation of secondary examination post resuscitation, while the highest compliance was initiating vasopressors if required (Supplementary Table 1, and Supplementary Fig. 1, There was no statistically significant difference between the compliant and noncompliant groups regarding the secondary outcomes of inhospital mortality, length of stay, or ED length of stay. Sixteen patients died (10%) in the hospital. The mortality between treatment that was compliant compared with noncompliant was 9% (n = 12) versus 11% (n = 4), respectively. Similarly, mortality from the year preceding the intervention (2017–2018) was 15%, which was not significantly different (p = 0.064), but a reduction in mortality of that magnitude may be clinically relevant. Overall sepsis bundle compliance was not significantly higher in patients who survived compared with those who died during hospitalization, 79% versus 75%, respectively (p = 0.273). The hospital length of stay was 5 days in noncompliant treatment on average compared with 6 days in those with compliant treatment (p = 0.474) (Fig. 2). The average ED length of stay was 377 minutes (median interquartile range [IQR], 279–440 min) in those with noncompliant treatment and 404 minutes (median IQR, 306–514 min) in those with complete compliance (p = 0.255) (Table 2). There was also no significant difference in the time of receiving the individual bundle components between compliant versus noncompliant treatment.

Table 2. - Time to Completion of Individual Components of Sepsis Bundles Between Patient Treatment That Was Complaint Versus Noncompliant
Time (min), Median (IQR) Sepsis Bundle Noncompliance (n = 35) Sepsis Bundle Compliance (n = 128) p
Time from triage to repeat lactic acid 148 (138–188) 156 (143–203) 0.39
Time from SRT paged to arrival 13 (10–16) 10 (8–15) 0.485
Time from triage to SRT at bedside 69 (52–170) 81 (52–121) 0.564
Time from triage to SRT paged 59 (42–171) 68 (41–106) 0.477
Time from triage to IVF 78 (16–148) 50 (18–99) 0.357
Time from ordering IVF to starting IVF 30 (2–36) 25 (1–36) 0.862
Time from ordering antibiotics to dose calculation 39 (21–50) 27 (10–49) 0.189
Time from triage to antibiotics ordered 87 (64–163) 80 (46–124) 0.142
Time from culture order to collection 10 (4–18) 8 (0–15) 0.218
Time from triage to culture collection 30 (16–46) 28 (18–54) 0.804
Time from triage to evaluation by an emergency department physician 8 (2–26) 9 (1–21) 0.685
Time from triage to point of care lactic acid time from triage 22 (17–88) 22 (15–40) 0.227
IQR = interquartile range, IVF = IV fluid, SRT = sepsis response team.

Figure 2.:
Impact of difference in compliance on overall length of stay and length of stay in emergency department as analyzed by fitting Cox proportional hazards regression models. SIRS = systemic inflammatory response syndrome.

A generalized linear model was used to identify variables that predicted sepsis compliance (Table 3). Variables identified as potential predictors included quick Sequential Organ Failure Assessment (SOFA) scores, lactic acid levels in mmol/L (< 2, ≥ 2, or ≥ 4), and admission or transfer to the ICU. Admission or transfer to the ICU was found to be a significant predictor of sepsis bundle compliance (95% CI, 0.035–0.309; p = 0.015). SOFA score was also collected and using a Fisher exact test; compliance did not vary significantly by SOFA score category (p = 0.905) (Supplementary Table 1, and Supplementary Fig. 1,

Table 3. - Individual Variables As Predictors of Sepsis Bundle Compliance As Determined by Generalized Linear Model
Predictor Estimate (95% CI) p
Quick Sequential Organ Failure Assessment score –0.051 (–0.134 to 0.032) 0.231
Lactic acid levels 0.048 (–0.044 to 0.139) 0.308
Admission or transfer to the ICU 0.172 (0.035–0.309) 0.015

Sepsis Bundle Compliance Before and After the Intervention of SRT

Following the completion of our SRT project, the sepsis bundle compliance appeared to rebound back to its historical range. A total of 18 out of 61 cases (30%) were compliant with all elements of the sepsis bundle. A one-sample t test of proportions shows that the compliance during the study period was significantly higher (p < 0.001).

Sepsis Bundle in the Coronavirus Disease 2019 Era

Of interest, given the unprecedented challenges inflicted by the coronavirus disease 2019 (COVID-19) pandemic, a retrospective look at sepsis bundle compliance during the peak of our COVID-19 cases was performed. Looking at the first 4 months of the COVID-19 crisis (March 2020 to June 2020), sepsis bundle compliance was 66% (57/87); further indicating that our study period had significantly higher compliance (p < 0.001).


Our study showed that a multifaceted QI project primarily led by the resident physicians of a major academic hospital led to a marked improvement in compliance with the sepsis bundles. The overall compliance was 79%, and the compliance across individual bundles ranged from 80% to 100%. We showed the feasibility of a resident physician-driven project that also involved hospital administration, ED nurses, ICU attending physicians, and ED attending physicians. It consisted of education of our house staff, enhanced communication between ED and ICU providers, structural change in how we respond to sepsis, and continuing assessment of the QI project as it evolved. We did not notice a significant change in clinical outcomes; however, the study was not powered to demonstrate a significant change in mortality or other outcomes.

As it becomes increasing clear that QI initiatives targeting improving screening measures and early recognition can lead to improved compliance with the sepsis bundles (10), it is important to assess the clinical benefit derived from better compliance with the individual bundle components. Timely antibiotic administration seems particularly important. In 2017, a large retrospective study in patients with sepsis found that a delay of bundle completion was associated with worsening mortality, as was a longer duration to receiving antibiotics (11). Conversely, a delay in fluid administration was not associated with increased mortality (11). Thus, not all bundle elements carry the same weight on patients’ outcomes. An emphasis has been placed on early antibiotic administration as each hour antibiotics are delayed has an associated increase in mortality, and this is most pronounced in patients with septic shock (12). This conglomeration of findings led to “The Surviving Sepsis Campaign Bundle: 2018 Update,” in which a recommendation of the hour-1 bundle to measure lactate, obtain cultures, administer antibiotics, begin fluid resuscitation, and if needed starting vasopressors (13). Our study was already in progress at that time, and it should be noted that except for mean time of antibiotic administration, which approached the goal, the targets for the hour-1 bundles on average were achieved.

The overall mortality between compliance and noncompliance with the resuscitation bundles was 9% and 11%, respectively. These numbers are lower than the inhospital mortality of 29% in high compliance and of 38.6% in low compliance patients with severe sepsis and septic shock in the trial published by Levy et al (9). This finding is likely influenced by the inclusion of patients with uncomplicated sepsis. The lack of a significant difference in mortality between the compliance and noncompliance groups could be due to small sample size and the fact that there was no difference in timely administration of antibiotics in these two groups. It is also important to recognize that compliance was higher in patients that were admitted or transferred to the ICU, indicating that critically ill patients that were more likely to benefit from these interventions also were more likely to receive all of the components of the sepsis bundles.

One strength of this study was the reinforcement of the principles of early identification and timely interventions occurring across multiple disciplines and departments. The addition of the SRT added a standardized approach to the management of sepsis.

The limitations of this study include that it was conducted at a single center, and therefore, applicability to other institutions should be taken into consideration. Also, the patient population in this study was relatively small and had a significantly lower associated mortality compared with prior studies.

The results of our study could be generalizable to patients presenting to the ED with sepsis at urban teaching hospitals. The results demonstrate the plausibility of implementing a resident-driven intervention to improve compliance with sepsis bundles. Further studies should be directed on a larger scale to determine if the findings of this study are reproducible and to assess potential benefits in terms of mortality and other clinical outcomes.


In conclusion, our study indicates that the implementation of a multidisciplinary approach to the treatment of sepsis in the form of a SRT significantly improves compliance by ensuring all elements of the sepsis bundle are administered in a timely manner. While we did not demonstrate a significant change in mortality, we were able to highlight a trend toward an improvement in mortality which may be clinically relevant.


This study was made possible by the combined efforts and contributions of the emergency department and ICU nurses and staff in addition to the house staff from the emergency medicine and internal medicine departments at University of Louisville Hospital.


1. Hershey TB, Kahn JM. State sepsis mandates - a new era for regulation of hospital quality. N Engl J Med. 2017; 376:2311–2313
2. Rivers E, Nguyen B, Havstad S, et al.; Early Goal-Directed Therapy Collaborative Group. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001; 345:1368–1377
3. Yealy DM, Kellum JA, Huang DT, et al.; ProCESS Investigators. A randomized trial of protocol-based care for early septic shock. N Engl J Med. 2014; 370:1683–1693
4. Peake SL, Delaney A, Bailey M, et al.; ARISE Investigators; ANZICS Clinical Trials Group. Goal-directed resuscitation for patients with early septic shock. N Engl J Med. 2014; 371:1496–1506
5. Mouncey PR, Osborn TM, Power GS, et al. Protocolised Management In Sepsis (ProMISe): A multicentre randomised controlled trial of the clinical effectiveness and cost-effectiveness of early, goal-directed, protocolised resuscitation for emerging septic shock. Health Technol Assess. 2015; 19:i–xxv, 1–150
6. Centers for Medicare & Medicaid Services. CMS Measures Inventory. Available at: Accessed March 27, 2018
7. Ferrer R, Martin-Loeches I, Phillips G, et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: Results from a guideline-based performance improvement program. Crit Care Med. 2014; 42:1749–1755
8. Schramm GE, Kashyap R, Mullon JJ, et al. Septic shock: A multidisciplinary response team and weekly feedback to clinicians improve the process of care and mortality. Crit Care Med. 2011; 39:252–258
9. Levy MM, Rhodes A, Phillips GS, et al. Surviving Sepsis Campaign: Association between performance metrics and outcomes in a 7.5-year study. Intensive Care Med. 2014; 40:1623–1633
10. Gatewood MO, Wemple M, Greco S, et al. A quality improvement project to improve early sepsis care in the emergency department. BMJ Qual Saf. 2015; 24:787–795
11. Seymour CW, Gesten F, Prescott HC, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med. 2017; 376:2235–2244
12. Liu VX, Fielding-Singh V, Greene JD, et al. The timing of early antibiotics and hospital mortality in sepsis. Am J Respir Crit Care Med. 2017; 196:856–863
13. Levy MM, Evans LE, Rhodes A. The Surviving Sepsis Campaign Bundle: 2018 update. Intensive Care Med. 2018; 44:925–928

critical care; emergency medicine; quality improvement; sepsis; sepsis bundles; sepsis response team

Supplemental Digital Content

Copyright © 2021 The Authors. Published by Wolters Kluwer Health, Inc. on behalf of the Society of Critical Care Medicine.