Sepsis is a life-threatening, medical emergency and a leading cause of inhospital mortality. The number of sepsis patients who are treated in the medicine ward range from 33% to 45% according to different authors, but all of them agree that a high proportion of patients are not initially treated in the ICUs.1,2 Non-ICU nurses should be able to recognize signs of early sepsis and organ dysfunction in order to trigger early treatment. In 1 study, 57.4% of rapid response team (RRT) patients had sepsis.3 Thus, RRT providers, including hospitalists and advanced practice providers (APPs), are responsible for recognizing and treating sepsis in more than half of all RRT activations. Timeliness of RRT notification in patients with sepsis is critical, as delays in sepsis recognition and RRT activation beyond 1 hour after meeting sepsis criteria are independently associated with increased 30-day mortality, hospital mortality, and hospital length of stay.4 Septic shock occurs late in the continuum with mortality exceeding 40%.5
One large multicenter sepsis quality improvement (QI) project conducted in 60 academic and community hospitals across the United States evaluated outcomes for patients with sepsis. Patients with sepsis transferring directly from the emergency department to the ICU experienced mortality of 26%, whereas patients admitted to inpatient units and later transferred to the ICU with sepsis experienced mortality of 40%.6 This highlights the importance of education for non-ICU nurses that includes the latest sepsis definitions and strategies for early recognition and timely, guideline-driven management.
The Centers for Medicare & Medicaid Services (CMS) implemented performance benchmarks to guide evidence-based sepsis management. These are time-sensitive indicators that at a minimum mandate obtaining blood cultures, administering antibiotics and fluid boluses, and measuring lactate.7 The CMS sepsis QI initiatives require organizations to demonstrate compliance with all of the established benchmarks to be eligible for reimbursement.
The traditional definition of sepsis currently used by CMS is listed in Supplemental Digital Content Table 1, available at: http://links.lww.com/JNCQ/A518, and is based on systemic inflammatory response criteria, an identified source of infection, and evidence of at least 1 organ dysfunction.7 A new definition of sepsis (Sepsis-3) was introduced in 2016, highlighting mortality risk by assigning an organ dysfunction score to each of 6 organ systems (Sequential [Sepsis-related] Organ Failure Assessment [SOFA] score).5 The new sepsis definition is as follows: “life-threatening organ dysfunction caused by a dysregulated host response to infection.”5 Patients with infection and evidence of organ dysfunction (higher SOFA scores) had higher mortality than patients with systemic inflammatory response alone.5 A change in the SOFA score of 2 or more is consistent with acute organ dysfunction and is associated with a mortality risk of 10% in hospitalized patients with presumed infection.5
While the SOFA organ dysfunction score is recommended for use in critically ill patients, the 2016 definitions also introduced use of a simplified assessment tool for non-ICU patients.5 The quick SOFA (qSOFA) criteria are recommended to assess non-ICU patients at risk for clinical deterioration.5 The qSOFA score consists of only 3 elements that are easily evaluated at the bedside: systolic blood pressure, respiratory rate, and level of consciousness (see Supplemental Digital Content Table 1, available at: http://links.lww.com/JNCQ/A518). One point is assigned for each element, and scores of 2 or higher constitute a positive qSOFA score. Use of qSOFA criteria had a strong predictive value for identifying patients at high risk for poor outcomes.5
A multidisciplinary sepsis team comprising physicians, APPs, nurses, and pharmacists at our hospital meets regularly to review data on sepsis patients and to evaluate compliance with national guidelines and CMS benchmarks. Overall, our monthly compliance with CMS measures averaged about 50% for a 6-month period, prompting discussion. One concern was that patients may be arriving to the ICU late in the sepsis continuum. Another concern was that nurses and providers were unfamiliar with the new sepsis definitions and screening tools.
The purpose of this project was to implement a widespread multidisciplinary sepsis education initiative for nurses, physicians, and APPs working with medical-surgical patients or responding to RRT calls. We hypothesized that targeting sepsis recognition using both qSOFA and organ dysfunction criteria and empowering nurses to trigger an RRT call based on those criteria would significantly impact time-to-sepsis recognition while bringing needed resources to the bedside for non-ICU patients. The premise was that improving time to recognition would subsequently shorten time to sepsis intervention for high-risk patients.
Setting and design
This QI project was conducted at a 446-bed tertiary care, level II trauma center. Approval was obtained from the hospital institutional review board as a quality initiative. A sepsis education program was developed and delivered to providers including nurses, hospitalists, and APPs for all medical-surgical inpatient areas and ICUs. To evaluate its impact, we conducted a pre- and posteducation retrospective descriptive evaluation of adult patients, 18 years of age and older, who met CMS criteria for the diagnosis of severe sepsis after transfer from the inpatient unit to an ICU (either as a result of RRT activation or independent decision to transfer). We excluded from evaluation patients admitted directly from the emergency department to the ICU or from an outlying facility to the ICU and patients who transitioned to comfort measures within 3 hours of meeting CMS sepsis criteria.
Patients meeting criteria for severe sepsis were identified from the hospital's preexisting sepsis database. Electronic records were then reviewed to identify 30 eligible patients for each of the cohorts. For the preimplementation cohort (March 2016 through March 2017), the project leader reviewed 192 patient records to identify 30 patients meeting criteria for inclusion. Following sepsis education, for the postimplementation cohort (May 2017 through December 2017), 181 records were reviewed to identify 30 patients meeting inclusion criteria.
Sepsis education initiative
The sepsis education program was developed for the hospital by the primary author and delivered to all providers by 2 nurses, increasing the consistency and fidelity of the intervention. This program targeted (1) early recognition of sepsis and use of qSOFA screening for risk assessment, (2) the importance of initiating prompt, time-sensitive treatments, (3) introduction to CMS sepsis performance benchmarks, (4) introduction to the new sepsis-3 definitions and goals of treatment, and (5) use of the existing electronic order set for sepsis treatment. The electronic order set is prepopulated and prompts the provider to order evidence-based, time-sensitive diagnostic studies and sepsis interventions. Use of the electronic order set is encouraged to facilitate compliance with the timing of blood cultures and sequential lactates (laboratory responds to all RRT calls), timely completion of fluid boluses, and effective communication of antibiotic urgency to the pharmacy. When fluids and antibiotics are ordered using the electronic order set, the order is then flagged for pharmacy as urgent because of suspected sepsis and prioritized to prompt immediate action. Orders placed outside of the order set do not necessarily trigger the same urgent response in the pharmacy queue.
All physicians and APPs on the hospitalist team, RRT members (nurses, ICU charge nurses, ICU physicians, and APPs), and nurses from each inpatient unit and ICU participated in the educational initiative. Staff from the emergency department, postanesthesia care unit, and outpatient and procedural units were excluded. Education was accomplished during April and May 2017, with a combination of lectures, slide presentations, and supplemental teaching materials. Staff were educated either formally at staff meetings or via informal presentations on each unit. Education was delivered by 2 sepsis education nurses who rounded on each unit and conducted short (average 10 minutes) educational sessions. Each nurse received a badge card that attached to his or her identification badge that summarized organ dysfunction criteria and the established hospital criteria for initiating a rapid response call, and this served as an ongoing reference following the education. The card also illustrated qSOFA criteria and a treatment prompt labeled “5 things in 5 minutes,” which summarized the targeted treatments to be accomplished for patients with potential sepsis. Education was delivered on 12 hospital units, and 1000 badge cards were distributed to the participants.
Interprofessional communication was also highlighted in the education. Nurses were instructed to specifically include terms that would trigger concern for sepsis and sepsis response when communicating changes in patient condition to physicians and APPs, for example, including phrases such as “My patient meets qSOFA criteria,” or “I am concerned about sepsis.” Nurses were empowered to independently activate RRT calls if qSOFA screening was positive with a score of 2 or more out of a possible 3 or to notify the physician or the APP of these changes in condition. Hospitalists and APPs were prompted to respond to these calls with urgency. Nurses were instructed that handoff communication between inpatient units and ICU should include the concern for sepsis and report on the initiation of diagnostic tests and timing of treatments to facilitate completion of fluid boluses and administration of antibiotics especially if not accomplished prior to transfer. This strategy was designed to ensure seamless care for the patients, regardless of locale. Once the sepsis education was completed, qSOFA screening was implemented hospital-wide for nurses to use as a common mechanism to prompt early identification of sepsis, initiation of an RRT call, and escalation of care (as needed) in collaboration with RRT responders.
Sepsis recognition and RRT time
Sepsis recognition was recorded by the project leader via review of electronic records to determine the time at which each patient met criteria for sepsis by the traditional CMS sepsis definition and met qSOFA criteria to determine whether patients at risk were identified earlier by use of the qSOFA screening. The triggers for positive qSOFA scores were recorded as well as the indication for initiating an RRT call (respiratory, neurologic, or hypotension). The timing of the RRT call in relation to hospital admission was recorded.
Time to treatment for sepsis interventions
We evaluated the time intervals between RRT call and time-sensitive sepsis interventions. If the patient was transferred electively to the ICU and not part of an RRT call, the time of ICU transfer was used in lieu of RRT time. The ICU transfer time was defined by the first documented ICU vital signs. Similarly, we evaluated time from sepsis recognition by both CMS criteria and qSOFA criteria to each sepsis intervention. The time-sensitive sepsis interventions included time to blood cultures, antibiotic administration, first lactate measurement, and completion of fluid bolus of 30 mL/kg, all reported in hours.
Compliance with time-sensitive sepsis interventions
We noted the percentages of patients in each group meeting each of the time-sensitive sepsis interventions from the time of sepsis identification by both CMS criteria and qSOFA criteria.
The project leader screened electronic records for eligibility and then collected data from each of the 60 patients, recording (1) demographics and comorbidities, length of stay, and discharge disposition; (2) the time at which each patient met criteria for sepsis by CMS sepsis definition and qSOFA criteria; (3) the time of RRT call and physiologic triggers for the RRT call; and (4) the times from RRT call and identification of sepsis to each treatment intervention (blood cultures, antibiotics, initial lactate, and fluid bolus).
Data were analyzed using IBM SPSS statistics version 24 (Armonk, New York). Categorical data were reported as frequency and percentage within group, and between-group comparisons were done using either χ2 test or Fisher exact test. Continuous data were reported as mean with standard deviation (age) or median with interquartile range, and comparisons between groups were done using either the Student t test (age) or the nonparametric Mann-Whitney U test.
There were no significant differences between the pre- and postimplementation groups in age, gender, comorbidities, or discharge disposition (see Supplemental Digital Content Table 2, available at: http://links.lww.com/JNCQ/A519). Seven patients (23%) in the pregroup and 10 patients (33%) in the postimplementation group died in the hospital. Median length of stay was significantly shorter in the postimplementation group by 4.8 days (pre: 13.6, post: 8.8; P = .025). More than 80% of patients in both groups had sepsis identified in clinical notes prior to the RRT call.
Sepsis recognition and RRT call
After the sepsis education initiative, there was a statistically significant improvement in median time to RRT for patients meeting qSOFA criteria from 11.8 hours pre [interquartile range: 3.4, 34.3] to 1.7 hours post [interquartile range: 0, 11.7], P = .005 (Table). The timing of the RRT in relation to admission did not differ significantly between groups, but there were differences in the physiologic trigger for the RRT call with significantly more patients with a respiratory event in the preimplementation group (83% vs 37% post, P < .001) and significantly more hypotension as trigger in the postimplementation group (13% pre vs 50% post, P = .002).
Time to treatment for sepsis interventions
Time to each sepsis intervention was recorded from RRT call and from the time of sepsis identification by both CMS criteria and qSOFA criteria (Table). In both groups, blood cultures were drawn before the RRT call, and there were no significant differences between groups in time to blood cultures. A significant reduction in time from qSOFA criteria to time of antibiotic administration was noted between the pre- and postimplementation groups (1.4 [−2.4, 6.2] hours pre vs −4.7 [−25.4, 1.8] hours [prior to meeting qSOFA criteria] post, P = .009) (Table). Time to fluid bolus administration from RRT call improved for both CMS and qSOFA groups, though it did not reach statistical significance for either group (Table).
Compliance with time-sensitive sepsis interventions
The percentages of patients in compliance with each of the time-sensitive treatments are noted in the Table. The definitions for assessing compliance are listed in Supplemental Digital Content Table 1, available at: http://links.lww.com/JNCQ/A518. There were no significant differences between groups in the proportion of patients having blood cultures within 1 hour of sepsis identification (either by CMS criteria or by qSOFA criteria), with all groups above 50% compliance. Compliance with antibiotic administration within 3 hours of sepsis identification improved significantly in the qSOFA group from 60% of patients in the preimplementation group to 87% of patients in the postimplementation group, P = .02. Compliance with completing the 30 mL/kg fluid bolus remained at less than 50% of patients in each group both pre and postimplementation. Of note, the CMS criteria can be met without hypotension, and compliance with fluid bolus required a full 30 mL/kg bolus. Compliance with first lactate measure remained low in both groups as well.
The high number of patients with a diagnosis of sepsis prior to the RRT (87% pre and 83% post) is an important finding as it reinforces the need for providing sepsis education to non-ICU nurses to promote early identification and early treatment of these high-risk patients. Educating on qSOFA and empowering nurses to call RRT based on qSOFA-positive scores were associated with shorter time to RRT and possibly shortened the time to intervention after the sepsis education initiative.
A recent retrospective Australian study by LeGuen et al8 conducted in a large tertiary hospital found that nearly 40% of RRT patients were qSOFA positive and subsequently had a significant difference in rates to ICU admission, repeat RRT, and inhospital death compared with infected patients who were qSOFA negative. This study reinforces the relevance of evaluating patients with qSOFA-positive scores. LeGuen et al8 limited their evaluation to validating qSOFA after RRT was called whereas our education focus was to use qSOFA as the activation criteria for RRT calls. Our results suggest that using qSOFA to trigger an RRT call is a valid approach to get help to the bedside and to initiate prompt intervention (blood cultures, lactate measures, antibiotic use, and fluid bolus) for sepsis patients in the medical ward at high risk for poor outcomes.
Delmas et al9 documented that a sepsis team performed superiorly with regard to intervention led by qSOFA in a retrospective analysis of about 1000 patients in a tertiary medical center. This reinforces our premise that qSOFA can be useful to lead early sepsis recognition. Our initiative differs from Delmas et al because in our hospital, the RRT team is in charge of responding to all inward emergencies, and we do not have a dedicated sepsis team. However, both results suggest an improvement in bundle compliance, which has been the expectation by CMS and other regulatory agencies.6
A different approach to early recognition for sepsis in the medicine ward is electronic sniffers such as the one reported by Guirgis et al.10 They propose the use of automated electronic health record data surveillance applying Modified Early Warning Signs-Sepsis Recognition Score (MEWS-SRS) that showed promising impact on early intervention.10 We had a comparable improvement in sepsis recognition when teaching nurses to use qSOFA to activate RRT calls (the postimplementation group decreased their RRT activation time almost by 9 hours in our small sample). The negative aspect related to electronic tools includes alarm fatigue and cost of implementation. Our objective was to cost-effectively impact recognition on septic patients with the resources available at our hospital. Implementation of electronic warning systems comes with higher cost. We acknowledge the limitation of an education-only intervention with potential wear off effect and the need for ongoing education to sustain improvement.11 As part of this initiative, we are planning to add this education to the mandatory yearly education package for nurses. Our QI was limited to 1 intervention, education. Other strategies suggested by Yost et al11 such as reminders, feedbacks, and audits could be completed and provide leverage for our results to increase the impact of our intervention in the long term. As part of a continuing QI initiative, these interventions will be considered for future application.
The sample size in each group is small, limiting the strength of some associations and the ability to generalize results. This study is not powered to evaluate mortality outcomes but it is to serve as proof of concept for future efforts on QI. This project was conducted at a single tertiary level II trauma center, which may limit generalizability. Data were collected retrospectively, relying on existing documentation that may have limited the ability to specifically time the identification of sepsis.
A hospital-wide multidisciplinary sepsis education program targeting sepsis recognition using both qSOFA and organ dysfunction criteria empowered nurses to trigger an RRT call based on those criteria. Subsequently, time-to-sepsis recognition and RRT activation for patients meeting qSOFA criteria resulted in a shortened time to some sepsis interventions, most notably antibiotic administration. This nurse-driven education initiative provided an evidence-based platform for enhanced communication between disciplines, facilitating faster, criteria-based evaluation of patients by a multidisciplinary team. Patient-centered interventions (timing of RRT call and time to antibiotics) served as the measures of success for this project.
1. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemiology of severe sepsis
in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001;29(7):1303–1310.
2. Stoller J, Halpin L, Weis M, et al Epidemiology of severe sepsis
: 2008-2012. J Crit Care. 2016;31(1):58–62.
3. Cross G, Bilgrami I, Eastwood GM, et al The epidemiology of sepsis
during rapid response team reviews in a teaching hospital. Anaesth Intensive Care. 2015;43(2):193–198
4. Barwise AT, Thongprayoon C, Gajic O, Jensen J, Herasevich V, Pickering BW. Delayed rapid response team activation is associated with increased hospital mortality, morbidity, and length of stay in a tertiary care institution. Crit Care Med. 2016;44(1):54–63.
5. Singer M, Deutschman M, Seymour CW, et al The third international consensus definitions for sepsis
and septic shock (sepsis
-3). JAMA. 2016;315(8):801–810.
6. Schorr C, Odden A, Evans L, et al Implementation of a multicenter performance improvement program for early detection and treatment of severe sepsis
in general medical-surgical wards. J Hosp Med. 2016;11(suppl 1):S32–S39.
8. LeGuen M, Ballueer Y, McKay R, et al Frequency and significance of qSOFA criteria during adult rapid response team reviews: a prospective cohort study. Resuscitation. 2018;122:13–18.
9. Delmas T, Price J, Birdwell A, et al Impact of a sepsis
team on outcomes in patients with sepsis
on the wards. Chest. 2017;152(4 suppl):A400.
10. Guirgis FW, Jones L, Esma R, et al Managing sepsis
: electronic recognition, rapid response teams, and standardized care save lives. J Crit Care. 2017;40:296–302.
11. Yost J, Ganann R, Thompson D, et al The effectiveness of knowledge translation interventions for promoting evidence-informed decision-making among nurses in tertiary care: a systematic review and meta-analysis. Implement Sci. 2015;10(1):98.