Impact of the National Early Warning Score-based sepsis response system on hospital-onset sepsis in a tertiary hospital in South Korea

Dong-gon Hyun; Sohyeon Lee; Sunhui Choi; Jeongsuk Son; So-Hee Park; Sang-Bum Hong; Chae-Man Lim

doi:10.4266/acc.000625

Articles

Page Path: HOME > Acute Crit Care > Volume 40(2); 2025 > Article

Original Article Rapid response system Impact of the National Early Warning Score-based sepsis response system on hospital-onset sepsis in a tertiary hospital in South Korea: Acute and Critical Care 2025;40(2):186-196.
DOI: https://doi.org/10.4266/acc.000625
Published online: May 20, 2025

Dong-gon Hyun¹

, Sohyeon Lee²

, Sunhui Choi²

, Jeongsuk Son²

, So-Hee Park^1,2

, Sang-Bum Hong¹

, Chae-Man Lim¹

¹Department of Pulmonary and Critical Care Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Korea

²Medical Emergency Team, Asan Medical Center, Seoul, Korea

Correspondence to: Chae-Man Lim Department of Pulmonary and Critical Care Medicine, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Korea Tel: +82 2 3010 4710 Fax: +82-2-3010-6968 Email: cmlim@amc.seoul.kr

• Received: February 10, 2025 • Revised: March 10, 2025 • Accepted: March 18, 2025

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

763 Views
47 Download

prev next

Full Article

Download PDF

Abstract
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
KEY MESSAGES
NOTES
SUPPLEMENTARY MATERIALS
REFERENCES

Abstract

Background
The effectiveness of electronic medical record-based alert systems, response protocols for sepsis diagnosis, and treatment in hospitalized patients remains unclear. This study aimed to determine whether the introduction of an electronic medical record-based sepsis response protocol (SRP) along with a 24/7 operating rapid response system affects the prognosis for patients with hospital-onset sepsis.
Methods
In August 2022, an SRP based on the National Early Warning Score was implemented in the electronic medical record system at Asan Medical Center. We retrospectively analyzed patients screened by the detection system for 1 year after the SRP implementation. Patients of the first 6 months (preliminary group) and those of the second 6 months (SRP group) were matched 1:1 based on propensity scores. The primary outcome was 30-day mortality.
Results
Of the 608 hospitalized patients screened by the system, 176 were assigned to each group after 1:1 propensity score matching. Patients in the SRP group were significantly more likely to receive blood cultures (58.5%) compared with the preliminary group (45.5%) (P=0.019). The SRP group showed a lower 30-day mortality risk (hazard ratio, 0.56; 95% CI, 0.36–0.86; P=0.017) compared to the preliminary group. A restricted cubic spline curve showed that SRP survival benefit began to manifest after the first 4 months (P=0.036).
Conclusions
Alongside an existing rapid response system, the National Early Warning Score-based SRP in the electronic medical record reduced mortality for hospital-onset sepsis within 1 year.
Key Words: early diagnosis; electronic health records; hospital rapid response team

INTRODUCTION

Sepsis affected an estimated 670 people per 100,000 globally and caused approximately 20% of all deaths in 2017, making it a major global health priority [1]. Among the different sepsis types, hospital-onset sepsis is characterized by higher comorbidities, increased organ failure scores, and a greater overall mortality rate compared with community-onset sepsis [2,3]. Additionally, hospital-onset sepsis patients are less likely to receive sepsis bundle treatments, compared to community-onset [4]. Given that sepsis is a medical emergency, various screening tools have been introduced for hospital-onset sepsis patients [5]. One proposed sepsis screening tool is a digital alert system integrated into electronic medical records (EMRs) to support clinical decision-making [6,7].

Numerous studies have investigated the effectiveness of sepsis alert systems [8], with a recent systematic review of 36 studies finding that they have a positive impact [9]. However, the effectiveness of hospital-onset sepsis alert systems is still controversial due to the significant heterogeneity in the proportion of hospital-onset sepsis cases across studies. Research has primarily focused on the early detection capabilities of alert systems (upstream) rather than the automatic activation of response protocols (downstream) [10-12]. Additionally, many hospitals utilize various types of rapid response teams (RRTs) to manage at-risk patients promptly [13-15]. Since RRT practices can significantly influence the outcomes of hospital-onset sepsis patients, it is crucial to assess the effectiveness of the sepsis response protocol (SRP) alongside a fully operational RRT [16]. This study aimed to evaluate the impact of an EMR-embedded automatic alert system with an SRP on clinical outcomes in hospital-onset sepsis patients supported by a 24/7 RRT.

MATERIALS AND METHODS

The Institutional Review Board of Asan Medical Center approved the study protocol (No. 2024-1153), and the requirement for obtaining patient informed consent was waived due to the retrospective design. The study was conducted according to the Guidelines for Good Clinical Practice and the Declaration of Helsinki.

Study Design

Asan Medical Center, a tertiary hospital with 2,700 beds and 193 critical care beds, has been running a 24-hour operating RRT comprising intensivists and highly trained nurses since March 2008. The center has introduced an SRP based on the National Early Warning Score (NEWS) system within the general ward’s EMR system using the fully operational RRT. This screening system has been operational for all general ward patients since August 2022.We conducted a retrospective analysis of adult hospitalized patients who recorded a NEWS of 7 or higher for the first time after admission, suspecting they had acquired sepsis during the year following implementation (August 2022–July 2023). We excluded patients who were in the intensive care unit, those who had a non-infectious cause for the NEWS increase, or those with missing clinical data.

SRP-Based Electric Medical Record System

Asan Medical Center employs the Asan Medical Information System version 3.0 as its EMR platform. Using this system, we developed an automated screening tool for early detection of sepsis based on the NEWS for all hospitalized patients. The early detection system is linked to the SRP, which sends an action plan to the attending physician. We chose the NEWS system over the quick sequential organ failure assessment score because of its superior accuracy in early sepsis detection [17-19].

The EMR system automatically calculates the NEWS in real time once the relevant data is entered during nursing assessments. The score is derived from various components assessed simultaneously by the attending nurse. The early detection system generates an “alert” pop-up window if the NEWS is 7 or higher. While all hospitalized patients are screened, alerts are only triggered for general ward patients. Once a nurse confirms the NEWS, the attending physician receives the pop-up window to verify the suspicion of sepsis (Figure 1). Based on the physician’s evaluation, the SRP is activated to assist clinicians in initiating the sepsis bundle, which comprises lactate measurement, blood culture, administration of antibiotics, fluid resuscitation, and vasopressor use, within one hour. If a patient reaches a NEWS of 7 or higher on the alert system and the activation criteria for the RRT are met (Supplementary Table 1), the team collaborates with attending physicians to implement the SRP.

Study Groups and Outcome Measures

To evaluate changes in clinical outcomes before and after implementation, we divided the study population into two groups: the preliminary group, which included patients screened and alerted by the system during the first 6 months post-implementation, and the SRP group, which included patients screened and alerted during the second 6 months. Propensity score matching was used to minimize the effect of confounding factors. The propensity score for the implementation period was estimated using a multivariable logistic regression model with sex, age, primary treatment department, and NEWS as covariates. Patients in the Preliminary and SRP groups were matched 1:1 using the nearest-neighbor algorithm without replacement and a caliper width of 0.1.

The primary outcome was 30-day mortality after NEWS alerts. We assessed the temporal relationship between system implementation and hospital mortality. Additionally, we evaluated the responses of nurses and physicians to the EMR notifications by examining: (1) the time the attending nurse responded to the pop-up window, (2) the time the attending physician closed the pop-up, (3) instances when the attending physician ordered the sepsis bundle, and (4) instances when the sepsis bundle was actually applied to the patients. We defined the time of the NEWS alerts as time zero. Compliance with the sepsis bundle protocol, excluding blood cultures and antibiotics, was considered successful if completed within 24 hours. Blood cultures and antibiotics were considered successful if completed within 48 hours before and after NEWS alerts.

Statistical Analysis

Data were presented as numbers and proportions for categorical variables and as means±standard deviations or medians (interquartile range [IQR]) for continuous variables with a normal or non-normal distribution. The distributions of how various components of the sepsis bundle were performed over time were illustrated using histograms. In the propensity score-matched cohort, absolute standardized mean differences (SMDs) were calculated to evaluate the imbalance between the groups before and after matching. For 30-day mortality, survival curves were calculated using the Kaplan-Meier method, while the hazard ratio (HR) was estimated using the Cox-proportional hazard regression model to compare outcomes. The results were presented as HR with a 95% CI. A restricted cubic spline analysis was conducted to evaluate the time-responsive relationship. Two-sided P-values of <0.05 denoted statistical significance. All analyses were performed using R software version 4.1.2 (R Core Team).

RESULTS

Patient Selection

We initially gathered data from 5,669 patients screened by the detection system between August 2022 and July 2023 (Figure 2). Of these, 5,061 patients were excluded, resulting in a final cohort of 608 patients. From this cohort, 329 (54.1%) were screened during the first 6 months, and the remaining 279 (45.9%) were screened in the second 6 months. Following 1:1 propensity score matching, 176 patients were assigned to the preliminary group and 176 patients were assigned to the SRP group. The differences in baseline characteristics were well-balanced in the matched cohort, with the SMD ≤0.1 (Supplementary Figure 1). Both groups had a similar proportion of males (62.5% vs. 65.9%, SMD=0.072) and the median ages were 68.0 (59.0–76.0) and 65.5 (58.0–74.0) for the preliminary and SRP groups, respectively (SMD=0.065) (Table 1). Most patients were admitted to the department of internal medicine (72.7% vs. 71.6%, SMD=0.024). There were no significant differences in the NEWS between the two groups and no significant differences in mental status, systolic blood pressure, diastolic blood pressure, mean blood pressure, pulse rate, respiratory rate, body temperature, peripheral oxygen saturation, and use of oxygen devices. The baseline characteristics of the entire cohort of 608 patients are presented in Supplementary Table 2.

Sepsis Response Protocol

There was no significant difference in the median times until nurses checked the alerts between the two groups (12.0 minutes [IQR, 3.0–30.0] for the preliminary group vs. 12.0 minutes [IQR, 3.0–28.0] for the SRP group, P=0.834) (Table 2). The median times until physicians checked the alerts were also similar (32.0 minutes [IQR, 13.0–86.5] vs. 33.0 minutes [IQR, 16.5–76.5], P=0.486). However, while compliance with each component of the sepsis bundle was similar in both groups, patients in the SRP group (58.5%) were more likely to receive blood cultures than those in the preliminary group (45.5%) (P=0.019). The proportion of patients receiving antibiotics within 6 h was 94.6 % in the SRP group and 88.3 % in the preliminary group (P=0.065). Histograms showing lactate, blood culture, and antibiotic performance over time revealed that most patients in both groups had lactate measurements taken after time zero, while blood cultures and antibiotics were often administered before time zero (Figure 3). The profile SRP profile for all patients is shown in Supplementary Table 3.

Outcomes

In total, 44 of 176 patients (25.0%) in the preliminary group and 27 of 176 patients (15.3%) in the SRP group died within 30 days of sepsis detection, revealing a significant difference in 30-day mortality between the two groups (HR, 0.56; 95% CI, 0.36–0.86; P=0.017) (Figure 4). The restricted cubic spline curve examining the time-dependent relationship between hospital mortality rates and the system’s implementation showed mortality rates began to decline after approximately 280 days (Figure 5).

More results on the timing relationships between the sepsis bundle components and hospital mortality are shown in Figure 6. Earlier prescription of blood cultures (P=0.010) and earlier blood collection for culture (P=0.008) were significantly associated with reduced hospital mortality. Moreover, hospital mortality rates tended to increase with time of antibiotic prescription (P=0.530) or antibiotic administration (P=0.572) away from NEWS alerts, although these findings were not statistically significant.

DISCUSSION

The NEWS-based SRP at our hospital in addition to a fully operating rapid response system resulted in reduced mortality for hospital-onset sepsis over the 1-year introduction period. Mortality began to gradually decrease 4 months after the system’s implementation. Sepsis bundles, including blood culture and antibiotics administration within six hours also showed improvements. Our results contribute to existing evidence showing that sepsis alert systems in hospitalized patients can potentially enhance responses to sepsis. This strength of this study lies in its analysis of changes in sepsis bundle completion and prognosis over time after the implementation of the system, unlike previous studies that only compared outcomes before and after the system’s introduction [9]. Among the bundle components, compliance with blood cultures, which was time-dependently associated with hospital mortality, increased after the alert system was implemented. Additionally, compliance with lactate measurement and antibiotic administration improved overall, although the differences were not significant because the rates were already high when the system was introduced. Since no other sepsis management programs were introduced during the study period, our alert system may have heightened physicians’ awareness in the general wards, facilitating protocol activation for clinically deteriorating patients, similar to findings reported by Na et al. [12]. The combination of these changes likely led to a decrease in mortality over time following the system’s introduction.

The rapid response system is a patient safety intervention used in hospitals to provide urgent care for patients experiencing clinical deterioration in general wards. Studies have consistently shown that a rapid response system reduces hospital mortality [20]. However, the effectiveness of the RRTs can vary due to differences in their structures [21]. For instance, a nationwide cohort study on hospital-onset sepsis found that all-day RRTs significantly reduced in-hospital mortality compared to non-all-day RRTs [16]. These variations in RRT intensity explain the conflicting results of previous studies on alert systems for hospital-onset sepsis [9]. Implementing an alert system with an automated activation protocol is also one modification aimed at improving outcomes [22]. A recent study reported that an automated alert program within a rapid response system improved overall outcomes [23]. Our current study demonstrated that (1) SRP, in addition to a fully operational RRT, decreased sepsis mortality, and (2) SRP directed primarily at general ward physicians was effective.

Previous studies on the impact of alert systems have mostly focused on patients admitted to the emergency room [24]. A recent systematic review and meta-analysis indicated that sepsis alert systems were associated with better outcomes and improved adherence to sepsis protocols, but only for adult patients with community-onset sepsis admitted to the emergency room [25]. By contrast, sepsis alert system benefits for hospital-onset sepsis have been less apparent [26]. For example, a single-center randomized controlled trial found no significant differences in new antibiotic orders or in-hospital mortality following the introduction of an EMR-generated alert system [11]. However, the differences in sepsis management between community-onset sepsis and hospital-onset sepsis may have influenced these findings [27]. In assessing the completion of the sepsis bundle, “time zero” is typically defined as the time of triage in the emergency room for community-onset sepsis [28]. However, defining time zero for hospital-onset sepsis is challenging due to the lack of a gold standard clinical definition or pathophysiologic signs [29]. This ambiguity is believed to contribute to low adherence to the sepsis bundle and uncertain associations between the bundle and clinical outcomes [4,30]. Our study replicates this phenomenon and suggests reasons for the lack of association between sepsis bundle administration and outcomes in hospital-onset sepsis. Unlike community-onset sepsis patients, who receive prompt treatment upon emergency room admission, hospital-onset sepsis occurs during ongoing management on hospital wards, complicating decisions about sepsis bundle components. In most cases in our study, blood cultures and antibiotic administrations were performed before NEWS alerts (time zero). Additionally, the heterogeneity of comorbidities, varying severity of underlying conditions, and high prevalence of multi-drug resistant organisms may also account for the inconsistent effects of the alert system-based sepsis bundle protocol for hospital-onset sepsis [29]. For example, recent studies suggest that early administration of antibiotics improved survival only in subgroups of sepsis patients, such as those with septic shock or metastatic malignancies [31,32]. Similarly, our study also found no significant association between earlier time-to-antibiotics and reduced mortality. Therefore, it is crucial to identify an appropriate “time zero” definition to establish accurate treatment practices for hospital-onset sepsis and to verify the effectiveness of the alert system with response protocol in future studies.

Some limitations are also worth mentioning. First, our study’s retrospective design limits our ability to draw definitive causal conclusions from our findings. Second, ours was a single-center study. In our institution, both the RRT and the alert system aid clinicians in managing acutely deteriorating patients. Different hospitals may have varying patient care systems, which could negatively affect the alert system’s effectiveness. Therefore, it is crucial to confirm our results in other healthcare settings. Finally, we did not assess potential unintended effects of the alert system, such as inappropriate antibiotics use or alert fatigue.

In conclusion, the NEWS-based alert system combined with an SRP integrated into an existing rapid response system was associated with reduced in-hospital mortality among patients with hospital-onset sepsis. The effectiveness of the alert system became evident about 4 months after its implementation.

KEY MESSAGES

▪ The implementation of a National Early Warning Score-based alert system combined with a sepsis response protocol at our hospital in addition to a fully operating rapid response system resulted in reduced mortality for hospital-onset sepsis over the 1-year introduction period.

▪ The effectiveness of the alert system became evident about 4 months after its implementation.

NOTES

CONFLICT OF INTEREST

No potential conflict of interest relevant to this article was reported.

FUNDING

None.

ACKNOWLEDGMENTS

None.

AUTHOR CONTRIBUTIONS

Conceptualization: DGH, SHP, CML. Methodology: DGH, SBH. Formal analysis: DGH, SHP, SBH, CL. Data curation: SL, SC, JS. Visualization: DGH, SL, SC, JS. Funding acquisition: DGH. Writing – original draft: DGH, CL. Writing – review & editing: SHP, SBH, CL.

SUPPLEMENTARY MATERIALS

Supplementary materials can be found via https://doi.org/10.4266/acc.000625.

Supplementary Figure 1.

Absolute standardized mean differences between the two groups before and after propensity score matching. The horizontal axis represents the standardized mean differences, the dashed line indicates the absolute standardized mean difference of 0.1. Open dots reflect values prior to matching, black dots reflect values after matching. Matching succeeded in reducing the standardized mean difference within an absolute value of 0.1. NEWS: National Early Warning Score.

acc-000625-Supplementary-Figure-1.pdf

Supplementary Table 1.

Rapid response team activation criteria

acc-000625-Supplementary-Table-1.pdf

Supplementary Table 2.

Baseline characteristics of all patients

acc-000625-Supplementary-Table-2.pdf

Supplementary Table 3.

Profile of NEWS-based sepsis response protocol of all patients

acc-000625-Supplementary-Table-3.pdf

Figure 1.

Workflow of the National Early Warning Score (NEWS)-based electronic medical record (EMR) system and Sepsis Response Protocol for early diagnosis of sepsis. RED: Rapid response, Emergency assessment, and Decision for escalation and ongoing monitoring; EKG: electrocardiogram; SpO₂: saturation of peripheral oxygen.

Figure 2.

Flowchart of analyzed patient cohorts. ICU: intensive care unit; SRP: sepsis response protocol.

Figure 3.

Histogram showing the intervals for each sepsis bundle component from time zero to completion. (A) Preliminary group. (B) Sepsis response protocol (SRP) group. The bars are spaced at 60-minute intervals.

Figure 4.

Kaplan-Meier estimates of cumulative probabilities of 28-day survival in the propensity score matched cohort. SRP: sepsis response protocol; HR: hazard ratio.

Figure 5.

The association between the timing of system implementation and hospital mortality. A restricted cubic spline model showing the hazard ratio for hospital deaths based on the number of days since the system’s implementation, adjusted for National Early Warning Score. The model included three knots at the 10th, 50th, and 90th percentiles of the days (with the 10th percentile as the reference). Solid lines stand for hazard ratios (HRs), shaded areas show 95% CIs, and dotted lines represent the knots.

Figure 6.

Association between the timing of prescription or completion of each bundle component and hospital mortality. The restricted cubic spline model shows the hazard ratio (HR) for hospital deaths based on the time after prescription or completion of bundle component. The model was conducted with three knots at the 10th, 50th, and 90th day percentiles. Reference is the 10th percentile in panels (A) and (B), the 50th percentile in panel (F), and the 90th percentile in panels (C), (D), and (E). Solid lines indicate the hazard ratios, shadowed shapes show 95% CIs, and dotted lines represent the knots.

Table 1.

Baseline characteristics of the matched cohort

Characteristic	Preliminary (n=176)	SRP (n=176)	SMD
Male	110 (62.5)	116 (65.9)	0.072
Age (yr)	68 (59–76)	66 (58–74)	0.065
Department
Internal medicine	128 (72.7)	126 (71.6)	0.024
General surgery	26 (14.8)	28 (15.9)	0.029
Obstetric gynecology	6 (3.4)	5 (2.8)	0.039
Neurology/neurosurgery	5 (2.8)	8 (4.5)	0.088
Others	11 (6.2)	9 (5.1)	0.042
NEWS score	7.0 (7.0–8.0)	7.0 (7.0–8.0)	0.028
Alert mental status	117 (66.5)	113 (64.2)	0.049
Systolic blood pressure (mm Hg)	118.0 (100.5–137.0)	119.0 (101.0–138.0)	0.015
Diastolic blood pressure (mm Hg)	71.5 (63.0–82.0)	70.5 (61.0–83.5)	0.007
Mean blood pressure (mm Hg)	88.0 (76.0–100.0)	88.0 (75.0–101.5)	0.011
Pulse rate per minute	111.6±18.1	111.6±19.7	0.003
Respiratory rate per minute	22.0 (19.5–26.0)	22.0 (19.5–25.0)	0.003
Body temperature (℃)	38.0 (37.0–38.8)	38.1 (37.0–38.8)	0.037
Peripheral oxygen saturation (%)	95.0 (92.5–97.0)	95.0 (92.0–97.0)	0.016
Use of oxygen device (%)	88 (50.0)	83 (47.2)	0.057
ICU admission	4 (2.3)	5 (2.8)	0.036

Values are presented as number (%), median (interquartile range), or mean±standard deviation.

SRP: sepsis response protocol; SMD: standardized mean difference; NEWS: National Early Warning Score; ICU: intensive care unit.

Table 2.

Profile of NEWS-based sepsis response protocol

Characteristic	Preliminary (n=176)	SRP (n=176)	P-value
Time interval until nurse checks (min)	12.0 (3.0 to 30.0)	12.0 (3.0 to 28.0)	0.834
Time interval until physician checks (min)	32.0 (13.0 to 86.5)	33.0 (16.5 to 76.5)	0.486
Lactate	132 (75.0)	142 (80.7)	0.248
Compliance within 1 hour	45 (34.1)	42 (29.6)	0.502
Time interval from time zero to order	27.0 (11.0 to 86.5)	28.0 (12.0 to 68.0)	0.809
Time interval from time zero to perform	86.0 (46.5 to 130.0)	81.5 (49.0 to 146.0)	0.764
Lactate value	1.8 (1.3 to 2.6)	1.8 (1.3 to 2.6)	0.762
Blood culture	80 (45.5)	103 (58.5)	0.019
Compliance within 24 hours	80 (100.0)	103 (100.0)	1.000
Time interval from time zero to order	–90.5 (–575.0 to 15.5)	18.0 (–219.0 to 54.5)	0.007
Time interval from time zero to perform	–51.5 (–443.0 to 53.0)	35.0 (–191.0 to 87.0)	0.011
Bacteremia	13 (16.2)	16 (15.5)	1.000
Antibiotic	163 (92.6)	167 (94.9)	0.509
Compliance within 1 hour	46 (28.2)	44 (26.3)	0.796
Compliance within 3 hours	118 (72.4)	120 (71.9)	1.000
Compliance within 6 hours	144 (88.3)	158 (94.6)	0.065
Time interval from time zero to order	–311.0 (–1,144.0 to 10.0)	–135.0 (–1,008.5 to 21.0)	0.051
Time interval from time zero to perform	24.0 (–120.0 to 119.0)	60.0 (–120.0 to 128.0)	0.127
Fluid resuscitation	31 (17.6)	23 (13.1)	0.301
Compliance within 1 hour	15 (48.4)	9 (39.1)	0.689
Time interval from time zero to order	34.0 (12.0 to 203.0)	64.0 (29.0 to 154.0)	0.441
Time interval from time zero to perform	65.0 (19.0 to 223.5)	118.0 (42.0 to 156.5)	0.501
Vasopressor	21 (11.9)	23 (13.1)	0.487
Compliance within 1 hour	5 (23.8)	8 (50.0)	0.192
Time interval from time zero to order	112.4±132.1	73.8±96.5	0.332
Time interval from time zero to perform	180.1±116.8	103.1±83.5	0.032

Values are presented as median (interquartile range), number (%), or mean±standard deviation.

NEWS: National Early Warning Score; SRP: sepsis response protocol.

REFERENCES

1. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, Tsoi D, Kievlan DR, et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet 2020;395:200-11.Article PubMed PMC
2. Rhee C, Wang R, Zhang Z, Fram D, Kadri SS, Klompas M, et al. Epidemiology of hospital-onset versus community-onset sepsis in U.S. hospitals and association with mortality: a retrospective analysis using electronic clinical data. Crit Care Med 2019;47:1169-76.Article PubMed PMC
3. Rose N, Spoden M, Freytag A, Pletz M, Eckmanns T, Wedekind L, et al. Association between hospital onset of infection and outcomes in sepsis patients: a propensity score matched cohort study based on health claims data in Germany. Int J Med Microbiol 2023;313:151593.Article PubMed
4. Baghdadi JD, Wong MD, Uslan DZ, Bell D, Cunningham WE, Needleman J, et al. Adherence to the SEP-1 sepsis bundle in hospital-onset v. community-onset sepsis: a multicenter retrospective cohort study. J Gen Intern Med 2020;35:1153-60.Article PubMed PMC PDF
5. Goulden R, Hoyle MC, Monis J, Railton D, Riley V, Martin P, et al. qSOFA, SIRS and NEWS for predicting inhospital mortality and ICU admission in emergency admissions treated as sepsis. Emerg Med J 2018;35:345-9.Article PubMed
6. Narayanan N, Gross AK, Pintens M, Fee C, MacDougall C. Effect of an electronic medical record alert for severe sepsis among ED patients. Am J Emerg Med 2016;34:185-8.Article PubMed
7. McRee L, Thanavaro JL, Moore K, Goldsmith M, Pasvogel A. The impact of an electronic medical record surveillance program on outcomes for patients with sepsis. Heart Lung 2014;43:546-9.Article PubMed
8. Joshi M, Ashrafian H, Arora S, Khan S, Cooke G, Darzi A. Digital alerting and outcomes in patients with sepsis: systematic review and meta-analysis. J Med Internet Res 2019;21:e15166. Article PubMed PMC
9. Zhang Z, Chen L, Xu P, Wang Q, Zhang J, Chen K, et al. Effectiveness of automated alerting system compared to usual care for the management of sepsis. NPJ Digit Med 2022;5:101.Article PubMed PMC PDF
10. Manaktala S, Claypool SR. Evaluating the impact of a computerized surveillance algorithm and decision support system on sepsis mortality. J Am Med Inform Assoc 2017;24:88-95.Article PubMed PDF
11. Downing NL, Rolnick J, Poole SF, Hall E, Wessels AJ, Heidenreich P, et al. Electronic health record-based clinical decision support alert for severe sepsis: a randomised evaluation. BMJ Qual Saf 2019;28:762-8.Article PubMed PMC
12. Na SJ, Ko RE, Ko MG, Jeon K. Automated alert and activation of medical emergency team using early warning score. J Intensive Care 2021;9:73.Article PubMed PMC PDF
13. Hillman K, Chen J, Cretikos M, Bellomo R, Brown D, Doig G, et al. Introduction of the medical emergency team (MET) system: a cluster-randomised controlled trial. Lancet 2005;365:2091-7.Article PubMed
14. Jones DA, DeVita MA, Bellomo R. Rapid-response teams. N Engl J Med 2011;365:139-46.Article PubMed
15. Wood KA, Ranji SR, Ide B, Dracup K. Rapid response systems in adult academic medical centers. Jt Comm J Qual Patient Saf 2009;35:475-82, 437. Article PubMed PMC
16. Hyun DG, Lee SY, Ahn JH, Huh JW, Hong SB, Koh Y, et al. Mortality of patients with hospital-onset sepsis in hospitals with all-day and non-all-day rapid response teams: a prospective nationwide multicenter cohort study. Crit Care 2022;26:280.Article PubMed PMC PDF
17. Durr D, Niemi T, Despraz J, Tusgul S, Dami F, Akrour R, et al. National Early Warning Score (NEWS) Outperforms Quick Sepsis-Related Organ Failure (qSOFA) Score for early detection of sepsis in the emergency department. Antibiotics (Basel) 2022;11:1518.Article PubMed PMC
18. Usman OA, Usman AA, Ward MA. Comparison of SIRS, qSOFA, and NEWS for the early identification of sepsis in the emergency department. Am J Emerg Med 2019;37:1490-7.Article PubMed
19. Qiu X, Lei YP, Zhou RX. SIRS, SOFA, qSOFA, and NEWS in the diagnosis of sepsis and prediction of adverse outcomes: a systematic review and meta-analysis. Expert Rev Anti Infect Ther 2023;21:891-900.Article PubMed
20. Winters BD. Rapid response systems. Crit Care Clin 2024;40:583-98.Article PubMed
21. Lyons PG, Edelson DP, Churpek MM. Rapid response systems. Resuscitation 2018;128:191-7.Article PubMed PMC
22. Vandegrift MA, Granata R, Totten VY, Kellett J, Sebat F. Review of 20 years of continuous quality improvement of a rapid response system, at four institutions, to identify key process responsible for its success. Crit Care Explor 2021;3:e0448. Article PubMed PMC
23. Escobar GJ, Liu VX, Schuler A, Lawson B, Greene JD, Kipnis P. Automated identification of adults at risk for in-hospital clinical deterioration. N Engl J Med 2020;383:1951-60.Article PubMed PMC
24. Hwang MI, Bond WF, Powell ES. Sepsis alerts in emergency departments: a systematic review of accuracy and quality measure impact. West J Emerg Med 2020;21:1201-10.Article PubMed PMC
25. Kim HJ, Ko RE, Lim SY, Park S, Suh GY, Lee YJ. Sepsis alert systems, mortality, and adherence in emergency departments: a systematic review and meta-analysis. JAMA Netw Open 2024;7:e2422823. Article PubMed PMC
26. Alberto L, Marshall AP, Walker R, Aitken LM. Screening for sepsis in general hospitalized patients: a systematic review. J Hosp Infect 2017;96:305-15.Article PubMed
27. Tonai M, Shiraishi A, Karumai T, Endo A, Kobayashi H, Fushimi K, et al. Hospital-onset sepsis and community-onset sepsis in critical care units in Japan: a retrospective cohort study based on a Japanese administrative claims database. Crit Care 2022;26:136.Article PubMed PMC PDF
28. Seymour CW, Gesten F, Prescott HC, Friedrich ME, Iwashyna TJ, Phillips GS, et al. Time to treatment and mortality during mandated emergency care for sepsis. N Engl J Med 2017;376:2235-44.Article PubMed PMC
29. Ginestra JC, Coz Yataco AO, Dugar SP, Dettmer MR. Hospital-onset sepsis warrants expanded investigation and consideration as a unique clinical entity. Chest 2024;165:1421-30.Article PubMed PMC
30. Baghdadi JD, Brook RH, Uslan DZ, Needleman J, Bell DS, Cunningham WE, et al. Association of a care bundle for early sepsis management with mortality among patients with hospital-onset or community-onset sepsis. JAMA Intern Med 2020;180:707-16.Article PubMed PMC
31. Im Y, Kang D, Ko RE, Lee YJ, Lim SY, Park S, et al. Time-to-antibiotics and clinical outcomes in patients with sepsis and septic shock: a prospective nationwide multicenter cohort study. Crit Care 2022;26:19.Article PubMed PMC
32. Hechtman RK, Kipnis P, Cano J, Seelye S, Liu VX, Prescott HC. Heterogeneity of benefit from earlier time-to-antibiotics for sepsis. Am J Respir Crit Care Med 2024;209:852-60.Article PubMed PMC

Figure & Data

References

Citations

Citations to this article as recorded by

PubReader
ePub Link

Cite

CITE: export Copy Format; Close

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Impact of the National Early Warning Score-based sepsis response system on hospital-onset sepsis in a tertiary hospital in South Korea

Acute Crit Care. 2025;40(2):186-196. Published online May 20, 2025

DOI: https://doi.org/10.4266/acc.000625

XML Download

Figure

Related articles

Impact of the National Early Warning Score-based sepsis response system on hospital-onset sepsis in a tertiary hospital in South Korea

Figure 1. Workflow of the National Early Warning Score (NEWS)-based electronic medical record (EMR) system and Sepsis Response Protocol for early diagnosis of sepsis. RED: Rapid response, Emergency assessment, and Decision for escalation and ongoing monitoring; EKG: electrocardiogram; SpO2: saturation of peripheral oxygen.

Figure 2. Flowchart of analyzed patient cohorts. ICU: intensive care unit; SRP: sepsis response protocol.

Figure 3. Histogram showing the intervals for each sepsis bundle component from time zero to completion. (A) Preliminary group. (B) Sepsis response protocol (SRP) group. The bars are spaced at 60-minute intervals.

Figure 4. Kaplan-Meier estimates of cumulative probabilities of 28-day survival in the propensity score matched cohort. SRP: sepsis response protocol; HR: hazard ratio.

Figure 5. The association between the timing of system implementation and hospital mortality. A restricted cubic spline model showing the hazard ratio for hospital deaths based on the number of days since the system’s implementation, adjusted for National Early Warning Score. The model included three knots at the 10th, 50th, and 90th percentiles of the days (with the 10th percentile as the reference). Solid lines stand for hazard ratios (HRs), shaded areas show 95% CIs, and dotted lines represent the knots.

Figure 6. Association between the timing of prescription or completion of each bundle component and hospital mortality. The restricted cubic spline model shows the hazard ratio (HR) for hospital deaths based on the time after prescription or completion of bundle component. The model was conducted with three knots at the 10th, 50th, and 90th day percentiles. Reference is the 10th percentile in panels (A) and (B), the 50th percentile in panel (F), and the 90th percentile in panels (C), (D), and (E). Solid lines indicate the hazard ratios, shadowed shapes show 95% CIs, and dotted lines represent the knots.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Impact of the National Early Warning Score-based sepsis response system on hospital-onset sepsis in a tertiary hospital in South Korea

Characteristic	Preliminary (n=176)	SRP (n=176)	SMD
Male	110 (62.5)	116 (65.9)	0.072
Age (yr)	68 (59–76)	66 (58–74)	0.065
Department
Internal medicine	128 (72.7)	126 (71.6)	0.024
General surgery	26 (14.8)	28 (15.9)	0.029
Obstetric gynecology	6 (3.4)	5 (2.8)	0.039
Neurology/neurosurgery	5 (2.8)	8 (4.5)	0.088
Others	11 (6.2)	9 (5.1)	0.042
NEWS score	7.0 (7.0–8.0)	7.0 (7.0–8.0)	0.028
Alert mental status	117 (66.5)	113 (64.2)	0.049
Systolic blood pressure (mm Hg)	118.0 (100.5–137.0)	119.0 (101.0–138.0)	0.015
Diastolic blood pressure (mm Hg)	71.5 (63.0–82.0)	70.5 (61.0–83.5)	0.007
Mean blood pressure (mm Hg)	88.0 (76.0–100.0)	88.0 (75.0–101.5)	0.011
Pulse rate per minute	111.6±18.1	111.6±19.7	0.003
Respiratory rate per minute	22.0 (19.5–26.0)	22.0 (19.5–25.0)	0.003
Body temperature (℃)	38.0 (37.0–38.8)	38.1 (37.0–38.8)	0.037
Peripheral oxygen saturation (%)	95.0 (92.5–97.0)	95.0 (92.0–97.0)	0.016
Use of oxygen device (%)	88 (50.0)	83 (47.2)	0.057
ICU admission	4 (2.3)	5 (2.8)	0.036

Characteristic	Preliminary (n=176)	SRP (n=176)	P-value
Time interval until nurse checks (min)	12.0 (3.0 to 30.0)	12.0 (3.0 to 28.0)	0.834
Time interval until physician checks (min)	32.0 (13.0 to 86.5)	33.0 (16.5 to 76.5)	0.486
Lactate	132 (75.0)	142 (80.7)	0.248
Compliance within 1 hour	45 (34.1)	42 (29.6)	0.502
Time interval from time zero to order	27.0 (11.0 to 86.5)	28.0 (12.0 to 68.0)	0.809
Time interval from time zero to perform	86.0 (46.5 to 130.0)	81.5 (49.0 to 146.0)	0.764
Lactate value	1.8 (1.3 to 2.6)	1.8 (1.3 to 2.6)	0.762
Blood culture	80 (45.5)	103 (58.5)	0.019
Compliance within 24 hours	80 (100.0)	103 (100.0)	1.000
Time interval from time zero to order	–90.5 (–575.0 to 15.5)	18.0 (–219.0 to 54.5)	0.007
Time interval from time zero to perform	–51.5 (–443.0 to 53.0)	35.0 (–191.0 to 87.0)	0.011
Bacteremia	13 (16.2)	16 (15.5)	1.000
Antibiotic	163 (92.6)	167 (94.9)	0.509
Compliance within 1 hour	46 (28.2)	44 (26.3)	0.796
Compliance within 3 hours	118 (72.4)	120 (71.9)	1.000
Compliance within 6 hours	144 (88.3)	158 (94.6)	0.065
Time interval from time zero to order	–311.0 (–1,144.0 to 10.0)	–135.0 (–1,008.5 to 21.0)	0.051
Time interval from time zero to perform	24.0 (–120.0 to 119.0)	60.0 (–120.0 to 128.0)	0.127
Fluid resuscitation	31 (17.6)	23 (13.1)	0.301
Compliance within 1 hour	15 (48.4)	9 (39.1)	0.689
Time interval from time zero to order	34.0 (12.0 to 203.0)	64.0 (29.0 to 154.0)	0.441
Time interval from time zero to perform	65.0 (19.0 to 223.5)	118.0 (42.0 to 156.5)	0.501
Vasopressor	21 (11.9)	23 (13.1)	0.487
Compliance within 1 hour	5 (23.8)	8 (50.0)	0.192
Time interval from time zero to order	112.4±132.1	73.8±96.5	0.332
Time interval from time zero to perform	180.1±116.8	103.1±83.5	0.032

Table 1. Baseline characteristics of the matched cohort

Values are presented as number (%), median (interquartile range), or mean±standard deviation.

SRP: sepsis response protocol; SMD: standardized mean difference; NEWS: National Early Warning Score; ICU: intensive care unit.

Table 2. Profile of NEWS-based sepsis response protocol

Values are presented as median (interquartile range), number (%), or mean±standard deviation.

NEWS: National Early Warning Score; SRP: sepsis response protocol.