U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Hall KK, Shoemaker-Hunt S, Hoffman L, et al. Making Healthcare Safer III: A Critical Analysis of Existing and Emerging Patient Safety Practices [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2020 Mar.

Cover of Making Healthcare Safer III: A Critical Analysis of Existing and Emerging Patient Safety Practices

Making Healthcare Safer III: A Critical Analysis of Existing and Emerging Patient Safety Practices [Internet].

Show details

3Sepsis Recognition

, M.A. and , M.D., M.S.

Introduction

Sepsis has been a leading cause of hospitalization and death in U.S. healthcare settings for many years, and accounts for more hospital admissions and spending than any other condition.1 As a result, preventing, diagnosing, and treating sepsis effectively has been a focus of patient safety and public health in recent years. In this chapter, we discuss two patient safety practices that aim to identify signs of sepsis and septic shock as quickly as possible so that treatment can be started: manual screening tools and electronic patient monitoring systems (PMSs).

Screening tools are manually administered paper or electronic forms that guide clinicians through a set of criteria as they are assessing a patient. The screening process is administered either at a care transition (e.g., presentation at the emergency department [ED] or to emergency medical services [EMS]) or at regular intervals (e.g., the start of every nursing shift). Current evidence indicates that performance (sensitivity/specificity) of the tools varies, especially in the prehospital setting. Evidence for process measure improvement (i.e., time to initiation of treatment) was of moderate strength in both the hospital and prehospital setting. Evidence for outcome measure improvement (mortality, hospital length of stay [LOS], intensive care unit [ICU] transfer, and ICU LOS) was sparse but showed a trend toward improvement. More high-quality studies are needed in diverse settings to test the effects of sepsis screening tools.

Automated systems continuously monitor patient status, such as vital signs, and alert a clinician if criteria for possible sepsis are met. These systems are becoming more widespread, especially in hospitals, which have sophisticated technology infrastructures. While the studies were inconsistent, there appears to be evidence of moderate strength in the current literature for improvement in both process and outcome measures for PMSs. More high-quality studies are needed to confirm these findings, and to identify implementation best practices and lessons learned.

Importance of Harm Area

Sepsis is a syndrome of life-threatening organ dysfunction due to a person’s systemic dysregulated response to infection.2 Sepsis can be caused by many types of infection (bacterial, fungal, and viral) and can affect any age group, from neonatal to geriatric. It is a common reason for hospital admission and readmission, with an estimated incidence of 6 percent of all hospital admissions, or more than 1 million admissions in the United States every year.3,4 Sepsis also has one of the highest mortality rates of any hospital condition, estimated at 15–30 percent.4,5 Tracking incidence and mortality over time is challenging due to shifting definitions and an increasing awareness of sepsis. Some studies show an increase in incidence and a decrease in mortality in recent years, but some show no significant change in either.4,6 Among subgroups, older adults and nursing home residents are much more likely to develop and die from sepsis compared with younger adults and non-nursing home residents.7 In 2013, $24 billion was spent treating sepsis, more than any other condition treated in U.S. hospitals.1

The symptoms of sepsis (e.g., high temperature, high blood pressure) are shared by many other conditions, making sepsis difficult to diagnose, especially in the early stages.8 In addition, sepsis can start suddenly and quickly lead to organ dysfunction and death.8 In response to this, international organizations such as the Society for Critical Care Medicine have focused on addressing the two problems that sepsis presents: delay in recognition and diagnosis of sepsis, and delay in start of treatment, which combined contribute to the high mortality rate for sepsis.9

The need for early recognition and rapid treatment have led to guidelines about how to treat septic patients, with aggressive interventions and timeframes. The most commonly adopted of these is the Surviving Sepsis Campaign (SSC) bundle, which has gone through many iterations, and includes starting broad-spectrum antibiotics and intravenous (IV) fluids, and obtaining blood culture and lactate measurements within a 1- to 6-hour timeframe.10 Many government agencies across the world have proposed measuring and evaluating hospital compliance with the bundle elements to strongly encourage its use. Most notably, since October 2015, the Centers for Medicare & Medicaid Services requires U.S. hospitals to report their performance on a composite process-of-care measure for severe sepsis and septic shock, and ties reimbursement to the measure results. There is occasionally tension between the goals of antibiotic stewardship and sepsis guidelines, with the former focused on reducing inappropriate use of broad-spectrum antibiotics, and the latter requiring rapid and barrier-free initiation of broad-spectrum antibiotics.11 Clinicians sometimes perceive antibiotic stewardship goals as being purely restrictive, thereby creating tension in decisions about antibiotics; however, good antibiotic stewardship encompasses appropriate administration of antibiotics, including when there is clinical suspicion for severe sepsis or septic shock. In addition, many clinicians have apprehension about the IV fluid level due to the risk of fluid overload.12

The need to diagnose sepsis unambiguously and quickly has led to development of various diagnostic criteria. The signs and thresholds used in these criteria vary but always include at least one vital sign with abnormal thresholds (heart rate [HR], respiratory rate [RR], blood pressure [BP], temperature, etc.), and sometimes include clinical assessments (mental status, suspicion of infection) and laboratory results (lactate, creatinine). The most commonly used criteria are the qSOFA (quick Sequential Organ Failure Assessment), the NEWS (National Early Warning Score), and the increasingly abandoned SIRS (systemic inflammatory response syndrome) criteria.13

Patient Safety Practice (PSP) Selection

A literature search was conducted on six sepsis PSPs in three databases (CINAHL®, MEDLINE®, and Cochrane), and resulting abstracts were reviewed for relevance. Some identified sepsis PSPs (e.g., clinical decision support) spanned multiple harm areas and appear in cross-cutting chapters. One sepsis PSP about readily available antibiotics did not have enough information to warrant a review. The two remaining PSPs (screening tools and patient monitoring systems) are specific to sepsis and have enough evidence to support a review.

Borrowing from the “failure to rescue” literature, diagnostic and treatment processes for sepsis can be grouped into two phases, afferent and efferent, each containing its own related practices.14 Figure 3.1 below is a conceptual model related to sepsis. The focus of the PSPs contained in this chapter is the afferent phase: how clinicians and hospitals use diagnostic criteria to recognize sepsis quickly, using either manual screening or continuous electronic monitoring. Because of the changing criteria for sepsis, the PSPs do not compare the accuracy of the various diagnostic criteria but rather the effect of these strategies in clinical practice settings. The efferent phase, including treatment for sepsis, occurs after screening/surveillance and is outside the scope of this chapter.

Figure 3.1. Conceptual Model for Sepsis.

Figure 3.1

Conceptual Model for Sepsis.

References for Introduction

1.
Torio CM, Moore BJ. National Inpatient Hospital Costs: The Most Expensive Conditions by Payer, 2013: Statistical Brief #204. Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville (MD): Agency for Healthcare Research and Quality (US); 2006. [PubMed: 27359025]
2.
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). doi: 801–10.10.1001/jama.2016.0287. [PMC free article: PMC4968574] [PubMed: 26903338] [CrossRef]
3.
Pfuntner A, Wier LM, Stocks C. Most Frequent Conditions in U.S. Hospitals, 2010: Statistical Brief #148. Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville (MD): Agency for Healthcare Research and Quality (US); 2006. https://www​.hcup-us.ahrq​.gov/reports/statbriefs/sb148.jsp. [PubMed: 23534077]
4.
Rhee C, Dantes R, Epstein L, Murphy DJ, Seymour CW, Iwashyna TJ, et al. Incidence and yrends of sepsis in US hospitals using clinical vs caims data, 2009–2014. JAMA. 2017;318(13):1241–9. doi: 10.1001/jama.2017.13836. [PMC free article: PMC5710396] [PubMed: 28903154] [CrossRef]
5.
Hatfield KM, Dantes RB, Baggs J, Sapiano M. R. P, Fiore AE, Jernigan JA, et al. Assessing variability in hospital-level mortality among U.S. medicare beneficiaries with hospitalizations for severe sepsis and septic Shock. Crit Care Med. 2018;46(11):1753–60. doi: 10.1097/ccm.0000000000003324. [PMC free article: PMC6774245] [PubMed: 30024430] [CrossRef]
6.
Gaieski DF, Edwards JM, Kallan MJ, Carr BG. Benchmarking the incidence and mortality of severe sepsis in the United States. Crit Care Med. 2013;41(5):1167–74. doi: 10.1097/CCM.0b013e31827c09f8. [PubMed: 23442987] [CrossRef]
7.
Ginde AA, Moss M, Shapiro NI, Schwartz RS. Impact of older age and nursing home residence on clinical outcomes of US emergency department visits for severe sepsis. J Crit Care. 2013;28(5):606–11. doi: 10.1016/j.jcrc.2013.03.018. [PMC free article: PMC3770757] [PubMed: 23683561] [CrossRef]
8.
Dantes RB, Epstein L. Combatting sepsis: a public health perspective. Clin Infect Dis. 2018;67(8):1300–2. doi: 10.1093/cid/ciy342. [PMC free article: PMC6557150] [PubMed: 29846544] [CrossRef]
9.
Seymour CW, Kahn JM, Martin-Gill C, Callaway CW, Yealy DM, Scales D, et al. Delays from first medical contact to antibiotic administration for sepsis. Crit Care Med. 2017;45(5):759–65. doi: 10.1097/ccm.0000000000002264. [PMC free article: PMC6065262] [PubMed: 28234754] [CrossRef]
10.
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. 2017;45(3):486–552. doi: 10.1097/ccm.0000000000002255. [PubMed: 28098591] [CrossRef]
11.
Pulia MS, Redwood R, Sharp B. antimicrobial stewardship in the management of sepsis. Emerg Med Clin North Am. 2017;35(1):199–217. doi: 10.1016/j.emc.2016.09.007. [PubMed: 27908334] [CrossRef]
12.
Akhter M, Hallare M, Roontiva A, Stowell J. 154 fluid resuscitation of septic patients at risk for fluid overload. Ann Emerg Med. 2017;70(4):S61–S2. doi: 10.1016/j.annemergmed.2017.07.181. [CrossRef]
13.
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(6):345–9. doi: 10.1136/emermed-2017-207120. [PubMed: 29467173] [CrossRef]
14.
Devita MA, Bellomo R, Hillman K, Kellum J, Rotondi A, Teres D, et al. Findings of the first consensus conference on medical emergency teams. Crit Care Med. 2006;34(9):2463–78. doi: 10.1097/01.Ccm.0000235743.38172.6e. [PubMed: 16878033] [CrossRef]

3.1. Patient Safety Practice 1: Sepsis Screening Tools

3.1.1. Practice Description

Identifying signs of sepsis as early as possible is critical to averting organ failure and risk of death.1 However, sepsis does not have a simple diagnostic test or specific symptoms that unambiguously indicate onset. International organizations have developed diagnostic criteria and have recommended screening patients at risk of sepsis using these criteria.2 Manual paper or electronic tools guide clinicians through the criteria as they assess a patient. The screening process generally takes place either during a care transition (e.g., presentation at the ED or to EMS) or at regular intervals (e.g., the start of every nursing shift). A tool’s embedded logic determines if the patient is suspected of having sepsis. If so, the clinician must start treatment as quickly as possible, which has been shown to increase survival.3,4

3.1.2. Methods

Key Findings

  • Performance of screening tools varied widely, especially in the prehospital setting. More research is needed to determine the optimal variables and thresholds for a sepsis screening tool.
  • There was moderate evidence of process measure improvement in the hospital setting with screening, including time to antibiotics. Prehospital evidence was sparse but showed improvement as well.
  • Evidence for outcome measures (e.g., mortality, ICU LOS, ICU transfer) was sparse but showed a trend toward improvement, although the improvement was not always significant.
  • Higher quality studies in diverse settings are needed to test the effects of sepsis screening tools.

To answer the question, “Do sepsis screening tools improve patient outcomes?” three databases (CINAHL®, MEDLINE®, and Cochrane) were searched for “sepsis” and related synonyms, as well as “screening,” “algorithm,” “triage tool,” “Early Warning Score,” “early alert,” and other similar terms from 2008 to 2018. The initial search yielded 998 results; after duplicates were removed, 923 were screened for inclusion and 53 full-text articles were retrieved. Of those, 26 were selected for inclusion in this review. Articles were excluded if the outcomes were not relevant, the article was out of scope (including no quantitative results), or the study design was insufficiently described. Studies in which screening tool implementation was accompanied by other significant sepsis interventions (e.g., changes in antibiotic delivery) are considered in Section 3.3.

General methods for this report are described in the Methods section of the full report.

For this patient safety practice, a PRISMA flow diagram and evidence table, along with literature-search strategy and search-term details, are included in the report appendixes A through C.

3.1.3. Evidence Summary

A summary of key findings related to sepsis screening tools is located in the Key Findings box. The following section reviews the applicable studies in more depth, by measure type and setting.

Fifteen of the 26 studies examining the use of sepsis screening tools took place in a hospital setting, 10 took place in a prehospital setting, and 1 took place in a nursing home. Over 20 different screening tools that incorporate somewhat different diagnostic criteria were used in the 26 studies. The indicators and thresholds used to determine if a patient screens positive for sepsis also differed across tools. Vital signs (HR, RR, BP, temperature, etc.) were present in all tools; clinical assessments (mental status, suspicion of infection) were also common, while laboratory results (lactate, creatinine) were used in only a few tools due to the time it takes to run lab tests and receive results back. Many studies used diagnostic criteria developed by consensus-based professional organizations, such as the qSOFA, MEWS (Modified Early Warning Score), and the SIRS criteria, but some studies tested other indicators and thresholds.

3.1.3.1. Sensitivities/Specificities of Screening Tools

Diagnostic performance of various screening tools for sepsis was reported in 20 of the 26 studies. None reported process measures or outcomes other than diagnostic performance. Twelve studies were retrospective cohort analyses that assessed whether the screening tool would have identified or ruled out sepsis correctly. Such studies support validity testing of the tools but have a lower strength of evidence than prospective studies because they were not implemented in a clinical setting. Despite these limitations, it is important to have a high-performing tool that reliably identifies and rules out sepsis before testing its effect on processes or outcomes of care. The hospital was the setting in 11 studies, while 8 were focused on the prehospital setting, and 1 focused on the nursing home setting.

The studies each report some or all of the following performance metrics for screening tools: sensitivity, specificity, positive predictive value, negative predictive value, and area under the receiver operating curve. The most widely reported were sensitivity and specificity. When deciding on an acceptable level of sensitivity and specificity for a tool, it is important to consider where the tool is implemented and the processes surrounding its use. For example, in a prehospital setting (EMS) or nursing home, high sensitivity is usually valued over specificity because patients will be reevaluated at the hospital before treatment is started. In a hospital setting, high specificity is also important to reduce alert fatigue and unnecessary treatment.5

3.1.3.1.1. Prehospital and Nursing Home

The sensitivity and specificity of the prehospital and nursing home screening tools varied widely. Seven of the eight prehospital studies were retrospective and they were addressed in a 2016 systematic review by Smyth and colleagues that found low to very-low quality evidence for the accuracy of prehospital sepsis screening tools. The authors attributed this to lack of EMS personnel training about sepsis and the inaccuracy of using SIRS criteria alone.a They conclude that more validation studies are needed to determine the efficacy of prehospital sepsis screening tools.6 Hunter et al. (2016) was the only prospective study, and it produced the highest sensitivity of any prehospital screening tool (0.90). That tool was implemented with EMS personnel and was based on SIRS criteria and end tidal carbon dioxide (ETCO2) measurement. Specificity of the tool was 0.54.7 The only study in a nursing home setting was a retrospective analysis of five different sepsis screening tools, which had sensitivity ranging from 0.27 to 0.79 and specificity ranging from 0.69 to 0.93.5 The performance of the prehospital tools is summarized in Table 3.1.

Table 3.1. Sensitivities and Specificities of Prehospital Studies.

Table 3.1

Sensitivities and Specificities of Prehospital Studies.

3.1.3.1.2. Hospital

Performance of screening tools in the hospital setting was tested in 11 studies: 7 in the ED, 3 in medical and/or surgical wards, and 1 in a surgical ICU. In the ED setting, Goerlich and colleagues’ triage screening tool had the most balanced performance, with sensitivity of 0.85 and specificity of 0.78. The tool was prospectively implemented in the ED of a tertiary hospital and used standard vital signs and muscle oxygen saturation (StO2) to generate a cumulative screening score.8 The other prospective screening tool, used in the ED setting by Singer and colleagues, achieved a high specificity (0.82) but a low sensitivity (0.34). This tool was implemented in a suburban academic medical center ED and used SIRS criteria and lactate measurement.9 In medical and/or surgical wards, Gyang et al. reported on a highly sensitive (0.95) and specific (0.92) tool that was prospectively implemented in a 26-bed medical/surgical intermediate care unit based on SIRS criteria and suspicion of infection.10 MacQueen et al. also reported on a highly sensitive (1.00) and specific tool (0.88) implemented in a general surgical unit that used routinely collected vital signs.11 In the one surgical ICU study, Wawrose and colleagues found that a screening tool based on vital signs outperformed a more complex tool on sensitivity (0.75 vs. 0.45) while maintaining a high specificity (0.85).12 The performance of the hospital tools is summarized in Table 3.2.

Table 3.2. Sensitivities and Specificities of Hospital Studies.

Table 3.2

Sensitivities and Specificities of Hospital Studies.

3.1.3.2. Effect on Process Measures

Process measures for a sepsis screening tool were reported in five studies, two in a prehospital setting and three in a hospital setting. The tools used in the studies were not independently validated, but the studies target important process goals, including timely administration of antibiotics and fluids, that have been shown to improve outcomes in patients with sepsis.3,4 Time to antibiotic administration was reported in all five studies, while time to lactate measurement was reported in four, time to fluid administration in three, and blood culture draw was reported in one study.

3.1.3.2.1. Prehospital

Both prehospital studies showed that use of a sepsis screening tool affected process timeliness measures, although only one effect reached significance; these studies had sample sizes of less than 300 and a moderate risk of bias. Hunter et al. (2019) showed that EMS personnel using a sepsis screening tool decreased time to IV fluid administration, blood culture draw, lactate level draw, and administration of antibiotics compared with septic patients who were not screened. They attribute this effect to hospitals preparing staff and supplies for a septic patient arrival, and EMS staff gaining IV access and/or starting IV fluids before hospital arrival.13 Guerra and colleagues found a non-significant decrease in time to antibiotics (p=0.07) for septic patients who were identified by EMS personnel using a screening tool, compared with those not identified by EMS and did not find a significant effect on any other process measures of timeliness.14

3.1.3.2.2. Hospital

Among the hospital screening tools that were evaluated for their effect on care processes, one was implemented in the ED and two in the ICU. While the study designs varied, all three studies showed a significant decrease in time to antibiotic administration or an increase in compliance with the SSC time guideline for antibiotic administration. For example, Patocka and colleagues showed that mean time to antibiotics decreased by 21 percent (p= 0.0074) after the implementation of an ED triage screening tool in a 637-bed urban tertiary hospital.15 Rincon et al. used a tele-health approach for ICU sepsis screening across 10 hospitals and found that it increased compliance with the SSC antibiotic administration guideline from 55 percent to 74 percent (p= 0.001), as well as increasing compliance with the guideline for IV fluids from 23 percent to 70 percent (p = 0.001).16 A significant improvement in time to lactate measurement was also found in all three studies, in both the ED and the ICU.1517

3.1.3.3. Effect on Outcome Measures

The ultimate goal of a patient safety practice is to improve the patient outcomes. Three sepsis screening tools were studied prospectively and measured patient outcomes: one in the prehospital setting and two in the hospital setting. All three studies were observational in design and had low to moderately sized samples. The outcomes studied were mortality, ICU admissions rate, and ICU LOS. Attributing improvement in these outcomes to sepsis screening tools is difficult, however, because patients with sepsis are generally older, have multiple comorbidities, and may have advance directives for end-of-life care. In addition, reasons for ICU transfer and ICU LOS are multifactorial and not necessarily correlated with sepsis or the use of a screening tool.13

3.1.3.3.1. Prehospital

Hunter et al. (2018 was the only prehospital study that measured patient outcomes. This study involved an EMS screening tool with a subsequent alert to the hospital; it found a significant reduction in ICU admissions rate (33% with screening vs. 52% without screening, p=0.003), and a non-significant reduction in mortality (11% with screening, 14% without screening, p=0.565).13

3.1.3.3.2. Hospital

In the hospital setting, one study focused on the ICU and one on the ED. Tedesco and colleagues found that a nurse-administered screening tool in the ED of a 320-bed community hospital led to a significant reduction in mortality (18.4% vs. 13.2% days; P = 0.015).18 Larosa and colleagues implemented an ICU sepsis screening tool in a 673-bed urban teaching hospital and found a significant reduction in mortality after controlling for factors such as mortality in emergency department sepsis (MEDS) score, leucopenia, and age (p=0.01). However, the sample size for this study was quite small (n=58).17

3.1.4. Implementation

Despite the lack of conclusive evidence of effectiveness, use of tools to screen patients for signs of sepsis is widespread due to the urgency for identifying sepsis, and based on guidelines and hospital quality performance measures. However, implementing these tools can prove challenging in terms of resource use and workflow change for staff.

3.1.4.1. Facilitators

Two common facilitators mentioned across studies were education of the clinical staff who will be responsible for administering the screening, and a tool that is easy to learn and use. First, educating nurses and EMS staff about sepsis pathophysiology helps them to better understand and interpret screening parameters, just as these staff are trained to recognize signs of stroke or cardiac arrest.19 This education may have the additional effect of increasing sepsis care quality, independent of the screening tool itself. Authors stressed that screening tools cannot substitute for the clinical acumen of staff.10 Second, a tool should be as easy as possible to fit into a clinician’s workflow, such as a checklist using a selected number of readily available or routinely collected variables.20 As a result, lab test results were generally excluded from screening tools. However, it is important to balance the simplicity of a tool and its ease of use with strong sensitivity and specificity. Other facilitators mentioned in these studies included consistent and complete documentation of vital signs on which screening algorithms are based, and standardized use of the tool across hospital units to reduce confusion and communication breakdowns when patients or staff move between units.5,21

3.1.4.2. Barriers

Screening every patient for signs of sepsis on a regular basis is labor and time intensive, regardless of the setting. The yield in terms of identifying emerging sepsis may also be low, depending on the prevalence of sepsis in the setting in question. Additionally, the frequency of screening (for example, once per hospital shift) can delay diagnosis of sepsis, defeating the purpose of the screening tool. As a result, transitions of care such as EMS ambulance transport and ED admission are often targeted as optimal times for screening.22,23 Other potential barriers include alert fatigue if the tool used is not specific enough, and a possible increase in drug resistance from more and longer use of antibiotics. However, there is no reported evidence about these effects. Finally, without proper training and an easy-to-use tool, adherence by clinical staff may be suboptimal, as reported by O’Shaughnessy et al., diminishing potential benefits.19

3.1.5. Resources

3.1.6. Gaps and Future Directions

It is clear from the available literature that higher quality studies (e.g., robust prospective, randomized, quasi-experimental) with larger sample sizes and diverse settings would quantify the effects of sepsis screening tools on process and outcome measures. In addition, the optimal set of variables and thresholds for rapidly identifying a septic patient is not completely settled.

With the emergence of automated electronic screening (see Section 3.2), the use of paper screening tools may be less common in the hospital setting, and more appropriate for prehospital settings such as EMS, nursing home, and home health. Robust studies on the effects of screening tools in these settings would be beneficial.

References for Section 3.1

1.
Alberto L, Marshall A P, Walker R, Aitken LM. Screening for sepsis in general hospitalized patients: a systematic review. J of Hosp Infect. 2017;96(4):305–15. doi: 10.1016/j.jhin.2017.05.005. [PubMed: 28506711] [CrossRef]
2.
Levy MM, Evans LE, Rhodes A. The Surviving surviving sepsis campaign bundle: 2018 update. J Intensive Care Medicine. 2018;44(6):925–8. doi: 10.1007/s00134-018-5085-0. [PubMed: 29675566] [CrossRef]
3.
Seymour CW, Kahn JM, Martin-Gill C, Callaway CW, Yealy DM, Scales D, et al. Delays from first medical contact to antibiotic administration for sepsis. Crit Care Med. 2017;45(5):759–65. doi: 10.1097/ccm.0000000000002264. [PMC free article: PMC6065262] [PubMed: 28234754] [CrossRef]
4.
Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med. 2006;34(6):1589–96. doi: 10.1097/01.Ccm.0000217961.75225.E9. [PubMed: 16625125] [CrossRef]
5.
Sloane PD, Ward K, Weber DJ, Kistler CE, Brown B, Davis K, et al. Cansepsis be detected in the nursing home prior to the need for hospital transfer? J Am Med Dir Assoc. 2018;19(6):492–6.doit: e1.10.1016/j.jamda.2018.02.001. [PubMed: 29599052] [CrossRef]
6.
Smyth MA, Brace-McDonnell SJ, Perkins GD. Identification of adults with sepsis in the prehospital environment: a systematic review. BMJ Open. 2016;6(8):e011218–e. doi: 10.1136/bmjopen-2016-011218. [PMC free article: PMC4985978] [PubMed: 27496231] [CrossRef]
7.
Hunter CL, Silvestri S, Ralls G, Stone A, Walker A, Papa L. A prehospital screening tool utilizing end-tidal carbon dioxide predicts sepsis and severe sepsis. Am J Emerg Med. 2016;34(5):813–9. doi: 10.1016/j.ajem.2016.01.017. [PubMed: 26879597] [CrossRef]
8.
Goerlich CE, Wade CE, McCarthy JJ, Holcomb JB, Moore LJ. Validation of sepsis screening tool using StO2 in emergency department patients. J Surg Res. 2014;190(1):270–5. doi: 10.1016/j.jss.2014.03.020. [PubMed: 24713469] [CrossRef]
9.
Singer AJ, Taylor M, Domingo A, Ghazipura S, Khorasonchi A, Thode HC, et al. Diagnostic characteristics of a clinical screening tool in combination with measuring bedside lactate level in emergency department patients with suspected sepsis. Acad Emerg Med. 2014;21(8):853–7. doi: 10.1111/acem.12444. [PubMed: 25155163] [CrossRef]
10.
Gyang E, Shieh L, Forsey L, Maggio P. A nurse-driven screening tool for the early identification of sepsis in an intermediate care unit setting. J Hosp Med. 2015;10(2):97–103. doi: 10.1002/jhm.2291. [PMC free article: PMC4816455] [PubMed: 25425449] [CrossRef]
11.
MacQueen IT, Dawes AJ, Hadnott T, Strength K, Moran GJ, Holschneider C, et al. Use of a hospital-wide screening program for early detection of sepsis in general surgery patients. Am Surg. 2015;81(10):1074–9. pmid: 26463311. [PubMed: 26463311]
12.
Wawrose R, Baraniuk M, Standiford L, Wade C, Holcomb J, Moore L. Comparison of sepsis screening tools’ ability to detect sepsis accurately. Surg Infect (Larchmt). 2016;17(5):525–9. doi: 10.1089/sur.2015.069. [PubMed: 27447053] [CrossRef]
13.
Hunter CL, Silvestri S, Ralls G, Stone A, Walker A, Mangalat N, et al. Comparing quick sequential organ failure assessment scores to end-tidal carbon doxide as mortality predictors in prehospital patients with suspected sepsis. West J Emerg Med. 2018;19(3):446–51. doi: 10.5811/westjem.2018.1.35607. [PMC free article: PMC5942006] [PubMed: 29760838] [CrossRef]
14.
Guerra WF, Mayfield TR, Meyers MS, Clouatre AE, Riccio JC. Early detection and treatment of patients with severe sepsis by prehospital personnel. J Emerg Med. 2013;44(6):1116–25. doi: 10.1016/j.jemermed.2012.11.003. [PubMed: 23321295] [CrossRef]
15.
Patocka C, Turner J, Xue X, Segal E. Evaluation of an emergency department triage screening tool for suspected severe sepsis and septic shock. J Healthc Qual. 2014;36(1):52–61. doi: 10.1111/jhq.12055. [PubMed: 24372995] [CrossRef]
16.
Rincon TA, Bourke G, Seiver A. Standardizing sepsis screening and management via a tele-ICU program improves patient care. Telemed J E Health. 2011;17(7):560–4.doi: 10.1089/tmj.2010.0225. [PubMed: 21718115] [CrossRef]
17.
Larosa JA, Ahmad N, Feinberg M, Shah M, Dibrienza R, Studer S. The use of an early alert system to improve compliance with sepsis bundles and to assess impact on mortality. Crit Care Res Pract. 2012;2012:980369-. doi: 10.1155/2012/980369. [PMC free article: PMC3296210] [PubMed: 22461981] [CrossRef]
18.
Tedesco ER, Whiteman K, Heuston M, Swanson-Biearman B, Stephens K. Interprofessional collaboration to improve sepsis care and survival within a tertiary care emergency department. Journal Emerg Nurs. 2017;43(6):532–8. doi: 10.1016/j.jen.2017.04.014. [PubMed: 28550958] [CrossRef]
19.
O’Shaughnessy J. CNE SERIES. Early Sepsis Identification. Medsurg Nurs. 2017;26(4):248–52
20.
Bayer Ol, Schwarzkopf D, Stumme C, Stacke A, Hartog CS, Hohenstein C, et al. An early warning scoring system to identify septic patients in the prehospital setting: the PRESEP score. Acad Emerg Med. 2015;22(7):868–71. doi: 10.1111/acem.12707. [PubMed: 26113162] [CrossRef]
21.
Bansal SS, Pawar PW, Sawant AS, Tamhankar AS, Patil SR, Kasat GV. Predictive factors for fever and sepsis following percutaneous nephrolithotomy: A review of 580 patients. Urol Ann. 2017;9(3):230–3. doi: 10.4103/UA.UA_166_16. [PMC free article: PMC5532888] [PubMed: 28794587] [CrossRef]
22.
Filbin MR, Thorsen JE, Lynch J, Gillingham TD, Pasakarnis CL, Capp R, et al. Challenges and opportunities for emergency department sepsis screening at triage. Sci Rep. 2018;8(1):11059-.doi: 10.1038/s41598-018-29427-1. [PMC free article: PMC6056466] [PubMed: 30038408] [CrossRef]
23.
Fitzpatrick D, McKenna M, Rooney K, Beckett D, Pringle N. Improving the management and care of people with sepsis. Emerg Nurse. 2014;22(1):18–24. doi: 10.7748/en2014.04.22.1.18.e1294. [PubMed: 24689480] [CrossRef]
24.
McClelland G & Jones J. A pilot study exploring the accuracy of pre-hospital sepsis recognition in the North East Ambulance Service. J Paramedic Med. 2015; 7 (9):459–465. doi: 10.12968/jpar.2015.7.9.459 [CrossRef]
25.
Polito CC, Isakov A, Yancey AH 2nd, et al. Prehospital recognition of severe sepsis: development and validation of a novel EMS screening tool. Am J Emerg Med. 2015;33(9):1119–1125. doi:10.1016/j.ajem.2015.04.024. [PMC free article: PMC4562872] [PubMed: 26070235] [CrossRef]
26.
Shiuh T, Sweeney T, Rupp R, Davis B, Reed III J. An emergency medical services sepsis protocolith point-of-care lactate accurately identifies out-of-hospital patients with severe infection and sepsis. Ann Emer Med. 2012;60(4s):S44. doi: 10.1016/j.annemergmed.2012.06.097. [CrossRef]
27.
Wallgren U, Castrén M, Svensson A, Kurland L. Identification of the adult septic patient in the pre-hospital setting: a comparison of two screening tools and clinical judgment. Eur J Emerg Med. 2014;21. doi: 10.1097/MEJ.0000000000000084. [PubMed: 24080997] [CrossRef]
28.
Berger T, Green J, Horeczko T, et al. Shock index and early recognition of sepsis in the emergency department: pilot study. West J Emerg Med. 2013;14(2):168–174. doi:10.5811/westjem.2012.8.11546. [PMC free article: PMC3628475] [PubMed: 23599863] [CrossRef]
29.
Scott HF, Donoghue AJ, Gaieski DF, Marchese RF, Mistry RD. Effectiveness of physical exam signs for early detection of critical illness in pediatric systemic inflammatory response syndrome. BMC Emerg Med. 2014;14:24. doi:10.1186/1471-227X-14-24. [PMC free article: PMC4289256] [PubMed: 25407007] [CrossRef]
30.
Shapiro N, Wolfe R, Wright SB, Moore R, Bates DW. Who Needs a Blood Culture? A prospectively derived and validated prediction rule. J Emerg Med. 2008; 35(3): 255–264. doi: 10.1016/j.jemermed.2008.04.001. [PubMed: 18486413] [CrossRef]
31.
Tirotta D, Gambacorta M, La Regina M, Attardo T, Lo Gullo A, Panzone F, Mazzone A, Campanini M,Dentali F. Evaluation of the threshold value for the modified early warning score (MEWS) in medical septic patients: a secondary analysis of an Italian multicentric prospective cohort (SNOOPII study). QJM. 2017; 110(6): 369–373. doi: 10.1093/qjmed/hcw229 [PubMed: 28069905] [CrossRef]
32.
Hunter CL, Silvestri S, Stone A, Shaughnessy A, Miller S, Rodriguez A, Papa L. Prehospital sepsis alert notification decreases time to initiation of CMS sepsis core measures. Am Journal of Emer Med. 2019; 37(1): 114–117. doi: 10.1016/j.ajem.2018.09.034. [PubMed: 30269999] [CrossRef]
33.
Nannan Panday R, Minderhoud T, & Alam N, Nanayakkara P. Prognostic value of early warning scores in the emergency department (ED) and acute medical unit (AMU): A narrative review. Eur J Intern Med. 2017; 45: 20–31. doi: 10.1016/j.ejim.2017.09.027. [PubMed: 28993097] [CrossRef]
34.
Roney JK, Whitley BE, Maples JC, Futrell LS, Stunkard KA, Long JD. Modified early warning scoring (MEWS): evaluating the evidence for tool inclusion of sepsis screening criteria and impact on mortality and failure to rescue. J Clin Nurs. 2015; 24: 3343–3354. doi:10.1111/jocn.12952. [PubMed: 26780181] [CrossRef]
35.
Shetty AL, Brown T, Booth T, Van KL, Dor-Shiffer DE, Vaghasiya MR, Eccleston CE, Iredell J. Systemic inflammatory response syndrome-based severe sepsis screening algorithms in emergency department patients with suspected sepsis. Emer Med Australasia. 2016; 28: 287–294. doi: 10.1111/1742-6723.12578. [PubMed: 27073105] [CrossRef]

3.2. Patient Safety Practice 2: Sepsis Patient Monitoring Systems

3.2.1. Practice Description

Identifying signs of sepsis in a patient as early as possible is critical to averting organ failure and risk of death.1 However, sepsis does not have a simple diagnostic test or specific symptoms that unambiguously indicates onset. International organizations have developed diagnostic criteria and recommend screening patients at risk of sepsis using these criteria.2 Automated electronic patient monitoring (i.e., surveillance) for signs of emerging sepsis is becoming more widespread, especially in hospitals, which have sophisticated technology infrastructures. Such systems automatically and continuously monitor data from telemetry devices and/or electronic health record (EHR) entries, and alert a clinician if set criteria for sepsis are met. If, after evaluation, a clinician determines that the patient has sepsis, the clinician must start treatment immediately to reduce mortality and improve patient outcomes.2 The goal is to decrease the time to treatment initiation for sepsis, which has been shown to increase survival.3,4

3.2.2. Methods

Key Findings

  • There was moderate evidence of process measure improvement across multiple types of hospital units, and evidence was most consistent outside of the ICU.
  • Evidence for outcome measures (e.g., mortality, ICU LOS, ICU transfer) was mixed, but over half of the studies showed a significant improvement, and several showed an absolute improvement that did not reach statistical significance.
  • Higher quality studies are needed to test the effects of sepsis monitoring systems on process and outcome measures.

To answer the question, “Does continuous patient monitoring for sepsis improve patient outcomes?” three databases (CINAHL®, MEDLINE®, and Cochrane) were searched for “sepsis” and related synonyms, as well “monitoring,” “surveillance,” and other similar terms, from 2008 to 2018. Additional relevant articles from other sources were added as they were found. The initial search yielded 345 results; after duplicates were removed and additional articles added, 350 were screened for inclusion and 55 full-text articles were retrieved. Of those, 15 were selected for inclusion in this review. Articles were excluded if the outcomes were not relevant, the article was out of scope (including not quantitative), or study design was insufficiently described. Studies about PMS implementation that also included significant sepsis interventions (e.g., changes in antibiotic delivery) are considered in Section 3.3.

General methods for this report are described in the Methods section of the full report.

For this patient safety practice, a PRISMA flow diagram and evidence table, along with literature-search strategy and search-term details, are included in the report appendixes A through C.

3.2.3. Evidence Summary

A summary of key findings related to sepsis PMS is located in the Key Findings box. This section reviews applicable studies in more depth, by measure type (process and outcome) and setting. Please note that sensitivities and specificities of PMSs are not examined because the algorithms within PMSs that scan for sepsis can be constantly adjusted to fit the needs of the setting and optimize performance, as opposed to a static manual screening tool. Upon designing and implementing a sepsis PMS, the clinicians/administrators typically test the system performance and adjust variable thresholds to best balance speed, sensitivity, and specificity for their setting.

All included studies took place in the hospital setting: five in the ICU, five in the ED, three in general units, one in a telemetry unit, and one in multiple hospital units (ICU, pediatric ICU, and medical/surgical units).

3.2.3.1. Effect on Process Measures

While assessing PMSs for effects on outcome measures (e.g., mortality) is the ultimate goal of this PSP, it is also important to evaluate whether a PMS improves sepsis care processes. Process measures are typically based on evidence-based clinical recommendations, and an improvement in process measures would indicate that patients are receiving care that has been shown to lead to better outcomes. Processes that are commonly targeted for improvement are the timely administration of antibiotics, lactate measurement, blood culture draw, and fluid administration. One or more process measures for sepsis PMSs were reported in nine studies: four in the ED, three in the ICU, and two in noncritical care units. Studies had various designs, including two randomized controlled trials (RCTs), one quasi-experimental study, and six observational pre/post studies. In addition, four systematic reviews covered this topic to some degree. The most commonly reported process measure was time to antibiotic administration (n=8), followed by time to lactate measurement and blood culture draw (n=5 each), and time to fluid administration (n=3).

A systematic review by Warttig and colleagues, which included RCTs conducted in the ICU through September 2017, determined that there is very low-quality evidence for any improvement in time to antibiotic administration after implementation of a PMS, and none of the studies they reviewed showed a significant improvement.5 None of these studies reported on any other process measures. Three other systematic reviews (Despins, Makam et al., and Alberto et al.) included both non-RCT and non-ICU studies, and found mixed results on improvement in sepsis process measures. Despins searched for automated sepsis detection in the hospital setting from 2005 to 2015;6 Makam and colleagues searched for electronic sepsis systems through June 2014;7 and Alberto and colleagues searched for both continuous monitoring and intermittent monitoring through June 2016.1 Several studies these authors reviewed (all observational and all outside of the ICU) reported that PMSs significantly improved time to administration of antibiotics, lactate draw, blood culture draw, and/or fluid administration. For example, Narayanan and colleagues, after implementing a PMS monitoring vital signs in the ED of an academic medical center, found that average time to antibiotic administration decreased from 61.5 minutes to 29.0 minutes (p=<0.001).8 The authors of one systematic review hypothesized that PMSs in the ICU may not be as effective as those outside of the ICU because clinicians in the ICU are already vigilant for signs of patient deterioration, so a sepsis alert may be redundant, among other reasons.7

Of the six studies we reviewed that were published after the systematic reviews were conducted, five found a significant effect of a PMS on at least one process measure. Of these five, one was an RCT and the others were observational studies. An RCT in two ICU units with a total of 32 beds at an urban medical center (Shimabukuro et al.) found that patients with automated sepsis monitoring received antibiotics an average of 2.76 hours earlier than patients in the control group and had blood cultures drawn an average of 2.79 hours earlier than patients in the control group.9 Austrian et al. was the only new study that found no effect of a PMS on time to first lactate measurement or antibiotic administration prior to blood cultures. This study was conducted in the ED and urgent care units of an urban academic medical center;10 it was a pre/post observational study with control of possible cofounders, and the authors suggested that alert fatigue from a tool with low positive predictive value contributed to the lack of impact on process measures.

3.2.3.2. Effect on Outcome Measures

The patient outcomes in the studies of automated PMSs included mortality, ICU transfer rate, hospital LOS, and ICU LOS. Outcome measures for sepsis PMSs were reported in 12 studies: 3 in the ED, 5 in the ICU, 2 in general units, 1 in a telemetry unit, and 1 in multiple hospital units (ICU, PCU, and medical/surgical units). It is difficult to attribute effects on any of these measures, or lack thereof, to a PMS intervention, because many patients who develop sepsis are older, have multiple comorbidities, and may have advance directives for end-of-life care, all of which also affect the outcomes of interest. In addition, reasons for ICU transfer and ICU LOS are multifactorial and not necessarily correlated with sepsis or the PMS.11

Eight of the 12 studies found a significant effect of a sepsis PMS in improving at least one outcome measure, and others showed absolute, but not statistically significant, improvements. The studies that showed a significant improvement included two RCTs, one quasi-experimental study, and five observational studies. Six of the 12 studies that reported mortality showed a statistically significant decrease after implementing a PMS. For example, Manaktala and Claypool found a 41–53 percent drop in sepsis mortality (p = 0.03–0.06) after implementing a PMS in the three general units of a 941-bed tertiary teaching hospital.12 A study in nine neonatal ICUs across the United States showed a significant reduction in mortality (8.1% vs. 10.2%, p = 0.04) after implementing a neonatal sepsis PMS.13 Several studies showed an absolute reduction in mortality that was not statistically significant. For example, Hooper and colleagues conducted an RCT of a “listening application” that monitored patient vital signs in the 35-bed medical ICU of a large academic tertiary medical center, and found 14 percent mortality in the control group and 10 percent in the intervention group (p = 0.29).14

Nine studies reported on hospital LOS, and four found a significant effect of the sepsis PMS. For example, McCoy and Das found a 9.55-percent decrease in hospital LOS after the implementation of a machine learning-based PMS in multiple hospital units (ICU, PCU, and medical/surgical units) in a 242-bed regional community hospital.15 In contrast, Manaktala and Claypool, described above, showed a significant decrease in mortality but did not find a significant decrease in hospital LOS.12

Only one of the four studies (Jung et al.) that reported on ICU LOS found a significant effect from a PMS. This was an observational study of a PMS implemented in a 34-bed surgical ICU in a large academic medical center.16 The studies that found no effect on ICU LOS varied in setting, with one implemented in the ED, one in a medical ICU, and one in all noncritical care units.b One study attributed lack of impact on ICU LOS to a PMS with poor predictive value,10 and one credited the already vigilant ICU staff;14 the third was underpowered to detect modest changes in ICU LOS. Two studies reported on ICU transfer rate, and neither found a significant effect on this or any other outcome measure.10,17 Several studies that showed significant effects on process measures showed no significant effects on outcome measures; for example Umschied and colleagues.17

3.2.4. Implementation

An automated surveillance system is less time consuming for staff than manual screening for sepsis and alerts clinicians in near real time to a patient’s deteriorating condition, more quickly than most manual screening strategies. However, implementing an automated PMS for sepsis can be difficult technologically, financially, and in terms of workflow changes for staff. The studies we reviewed identified supporting factors that facilitate PMS implementation, as well as barriers to successful PMS implementation.

3.2.4.1. Facilitators

As with manual screening tools, implementing a PMS will be effective only if the system has a high level of sensitivity and specificity, to engender clinician trust and reduce false-positive alerts. To achieve this, some prospective studies iteratively revised thresholds for key values, with input from the clinicians, to optimize tool performance.15,18 Some more recent studies used machine learning to optimize system performance.9,18 To improve system usability, input from clinicians was solicited in some studies, followed by adaptations. These included allowing a nurse to “snooze” an alert for 6 hours if the patient is already under assessment for sepsis, or implementing a “traffic light” system on a dashboard to visually show clinicians which patients are in a warning zone (yellow) or need urgent attention (red).15,19 Other facilitators mentioned in the studies included: consistent and complete input of vital signs on which the PMS relies, having a specific staff member assigned to receive all alerts and determine if a physician needs to be called, and designing the PMS to work reliably even if data are incomplete.15,20,21 Building an automated PMS from scratch is costly, but several PMS systems are now available as an add-on EHR or telemedicine module, which is more efficient for a hospital than designing and testing a de novo system.

3.2.4.2. Barriers

The nonspecific nature of sepsis makes achieving a highly predictive system difficult, whether on paper or in an automated PMS. This is particularly difficult in pediatric settings because the “normal” ranges for vital signs are age dependent and more difficult to fine tune.22 In addition, if the electronic monitoring and alerting system is poorly designed or difficult to use, it can lead to clinician confusion, frustration, and possibly to worse patient care.23 For example, if the alert physicians receive contains too little information (or too much), or if the action required is not clear, physicians may find the system too difficult or burdensome to use.23,24 Lack of adequate staff training on using the system is also a potential barrier, even if a system has high sensitivity and specificity. Additionally, the cost of designing and implementing a PMS can be prohibitive for smaller hospitals, and while an EHR add-on can reduce cost, it may result in less customizable functionality. Finally, after a system is implemented, refining the algorithm and updating it based on changing sepsis criteria require close work with the facility’s IT department, which can be resource and time intensive.

3.2.5. Resources

The nonprofit Patient Safety Movement Foundation offers a toolkit on early sepsis detection that includes a technology plan for an automated PMS: http://patientsafetymovement.org/wp-content/uploads/2016/02/10-Sepsis-April-2016.pdf.

3.2.6. Gaps and Future Directions

Due to the mixed results, more high-quality studies could help to understand the effects of sepsis PMSs on important process and outcome measures in different hospital units.

The emergence of machine learning technology has the potential to improve the accuracy, consistency, and customizability of PMSs. Rather than rules-based patient monitoring with predetermined thresholds, machine learning can continually learn from sepsis and nonsepsis cases, and be able to better and more quickly predict when a patient is at risk of sepsis.15 More studies testing the effect of these systems on processes and outcomes are needed. In addition, the design and usability of systems could benefit from additional studies to determine the optimal display of alerts, dashboards, and other clinical decision support.

References for Section 3.2

1.
Alberto L, Marshall A P, Walker R, Aitken LM. Screening for sepsis in general hospitalized patients: a systematic review. J Hosp Infect. 2017;96(4):305–15.doi: 10.1016/j.jhin.2017.05.005. [PubMed: 28506711] [CrossRef]
2.
Levy MM, Evans LE, Rhodes A. The Surviving Sepsis Campaign Bundle: 2018 update. J Intensive Care Med. 2018;44(6):925–8. doi: 10.1007/s00134-018-5085-0. [PubMed: 29675566] [CrossRef]
3.
Seymour CW, Kahn JM, Martin-Gill C, Callaway CW, Yealy DM, Scales D, et al. Delays from first medical contact to antibiotic administration for sepsis. Crit Care Med. 2017;45(5):759–65.doi: 10.1097/ccm.0000000000002264. [PMC free article: PMC6065262] [PubMed: 28234754] [CrossRef]
4.
Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Crit Care Med. 2006;34(6):1589–96. doi: 10.1097/01.Ccm.0000217961.75225.E9. [PubMed: 16625125] [CrossRef]
5.
Warttig S, Alderson P, Evans DJ, Lewis SR, Kourbeti IS, Smith AF. Automated monitoring compared to standard care for the early detection of sepsis in critically ill patients. Cochrane Database Syst Rev. 2018;6:CD012404. doi: 10.1002/14651858.CD012404.pub2. [PMC free article: PMC6353245] [PubMed: 29938790] [CrossRef]
6.
Despins LA. Automated detection of sepsis using electronic medical record data: a systematic review. Journal Healthc Qual. 2017;39(6):322–33. doi: 10.1097/JHQ.0000000000000066. [PubMed: 27631709] [CrossRef]
7.
Makam AN, Nguyen OK, Auerbach AD. Diagnostic accuracy and effectiveness of automated electronic sepsis alert systems: a systematic review. J Hosp Med. 2015;10(6):396–402.doi: 10.1002/jhm.2347. [PMC free article: PMC4477829] [PubMed: 25758641] [CrossRef]
8.
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(2):185–8. doi: 10.1016/j.ajem.2015.10.005. [PubMed: 26573784] [CrossRef]
9.
Shimabukuro DW, Barton CW, Feldman MD, Mataraso SJ, Das R. Effect of a machine learning-based severe sepsis prediction algorithm on patient survival and hospital length of stay: a randomised clinical trial. BMJ Open Respir Res. 2017;4(1):e000234–e. doi: 10.1136/bmjresp-2017-000234. [PMC free article: PMC5687546] [PubMed: 29435343] [CrossRef]
10.
Austrian JS, Jamin CT, Doty GR, Blecker S. Impact of an emergency department electronic sepsis surveillance system on patient mortality and length of stay. JAMIA. 2018;25(5):523–9. doi: 10.1093/jamia/ocx072. [PMC free article: PMC7646853] [PubMed: 29025165] [CrossRef]
11.
Hunter CL, Silvestri S, Ralls G, Stone A, Walker A, Mangalat N, et al. Comparing quick sequential organ failure assessment scores to end-tidal Ccarbon dioxide as mortality predictors in prehospital patients with suspected sepsis. West J Emerg Med. 2018;19(3):446–51.doi: 10.5811/westjem.2018.1.35607. [PMC free article: PMC5942006] [PubMed: 29760838] [CrossRef]
12.
Manaktala S, Claypool SR. Evaluating the impact of a computerized surveillance algorithm and decision support system on sepsis mortality. JAMIA. 2017;24(1):88–95. doi: 10.1093/jamia/ocw056. [PMC free article: PMC7654083] [PubMed: 27225197] [CrossRef]
13.
Moorman JR, Delos JB, Flower AA, Cao H, Kovatchev BP, Richman JS, et al. Cardiovascular oscillations at the bedside: early diagnosis of neonatal sepsis using heart rate characteristics monitoring. Physiol Meas. 2011;32(11):1821–32. doi: 10.1088/0967-3334/32/11/S08. [PMC free article: PMC4898648] [PubMed: 22026974] [CrossRef]
14.
Hooper MH, Weavind L, Wheeler AP, Martin JB, Gowda SS, Semler MW, et al. Randomized trial of automated, electronic monitoring to facilitate early detection of sepsis in the intensive care unit. Crit Care Med. 2012;40(7):2096–101. doi: 10.1097/CCM.0b013e318250a887. [PMC free article: PMC4451061] [PubMed: 22584763] [CrossRef]
15.
McCoy A, Das R. Reducing patient mortality, length of stay and readmissions through machine learning-based sepsis prediction in the emergency department, intensive care unit and hospital floor units. BMJ Open Quality. 2017;6(2):e000158–e. doi: 10.1136/bmjoq-2017-000158. [PMC free article: PMC5699136] [PubMed: 29450295] [CrossRef]
16.
Jung AD, Droege CA, Johannigman J, Goodman M. Sooner is better: use of a real-time automated bedside dashboard improves sepsis care. J Surg Res. 2018; 231: 373–379. doi: 10.1016/j.jss.2018.05.078 [PubMed: 30278956] [CrossRef]
17.
Umscheid CA, Betesh J, VanZandbergen C, Hanish A, Tait G, Mikkelsen ME, et al. Development, implementation, and impact of an automated early warning and response system for sepsis. J Hosp Med. 2015;10(1):26–31. doi: 10.1002/jhm.2259. [PMC free article: PMC4410778] [PubMed: 25263548] [CrossRef]
18.
Amland RC, Hahn-Cover KE. Clinical decision support forearly recognition of sepsis. Am J Qual. 2016;31(2):103–10. doi: 10.1177/1062860614557636. [PMC free article: PMC4776220] [PubMed: 25385815] [CrossRef]
19.
Cruz AT, Williams EA, Graf JM, Perry AM, Harbin DE, Wuestner ER, et al. Test characteristics of an automated age- and temperature-adjusted tachycardia alert in pediatric septic shock. Pediatr Emerg Care. 2012;28(9):889–94.doi: 10.1097/PEC.0b013e318267a78a. [PubMed: 22929140] [CrossRef]
20.
Desautels T, Calvert J, Hoffman J, Jay M, Kerem Y, Shieh L, et al. Prediction of sepsis in the intensive care unit with minimal electronic health record data: a machine learning approach. JMIR Med Inform. 2016;4(3):e28–e. doi: 10.2196/medinform.5909, [PMC free article: PMC5065680] [PubMed: 27694098] [CrossRef]
21.
Mao Q, Jay M, Hoffman JL, Calvert J, Barton C, Shimabukuro D, et al. Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and ICU. BMJ Open. 2018;8(1):e017833–e. doi: 10.1136/bmjopen-2017-017833. [PMC free article: PMC5829820] [PubMed: 29374661] [CrossRef]
22.
Sepanski R J, Godambe SA, Mangum CD, Bovat CS, Zaritsky AL, Shah SH. Designing a pediatric severe sepsis screening tool. Front Pediatr. 2014;2:56. doi: 10.3389/fped.2014.00056. [PMC free article: PMC4058838] [PubMed: 24982852] [CrossRef]
23.
Pertiwi AAP, Fraczkowski D, Stogis SL, Lopez KD. Using heuristic evaluation to improve sepsis alert usability. Crit Care Nurs Clin North Am. 2018;30(2):297–309.doi: 10.1016/j.cnc.2018.02.011. [PubMed: 29724447] [CrossRef]
24.
Semler MW, Weavind L, Hooper MH, Rice TW, Gowda SS, Nadas A, et al. An electronic tool for the evaluation and treatment of sepsis in the ICU: a randomized controlled trial. Crit Care Med. 2015;43(8):1595–602. doi: 10.1097/CCM.0000000000001020. [PMC free article: PMC4506222] [PubMed: 25867906] [CrossRef]
25.
Berger T, Birnbaum A, Bijur P, et al. A computerized alert screening for severe sepsis in emergency department atients increases lactate testing but does not improve inpatient mortality. Appl Clin Inf. 2010; 1: 394–407. doi: 10.4338/ACI-2010-09-RA-0054. [PMC free article: PMC3633314] [PubMed: 23616849] [CrossRef]
26.
McRee L, Thanavaro J, Moore K, et al. The impact of an electronic medical record surveillance program on outcomes for patients with sepsis. Heart Lung. 2014; 43: 546–549. 10.1016/j.hrtlng.2014.05.009. [PubMed: 24961633] [CrossRef]
27.
Nelson J, Smith B, Jared J, et al. Prospective trial of real-time electronic surveillance to expedite early care of severe sepsis. Ann Emer Med. 2011; 57(5), 500–4. doi: 10.1016/j.annemergmed.2010.12.008. [PubMed: 21227543] [CrossRef]
28.
Sawyer AM, Deal EN, Labelle AJ, et al. Implementation of a real-time computerized sepsis alert in non-intensive care unit patients. Crit Care Med. 2011; 39(3):469–473. doi: 10.1097/CCM.0b013e318205df85. [PubMed: 21169824] [CrossRef]
29.
Croft CA, Moore FA, Marker PS, et al. Computer versus paper system for recognition and management of sepsis in surgical intensive care. J Trauma Acute Care Surg. 2014;76(2):311–7. doi: 10.1097/TA.0000000000000121. [PubMed: 24458039] [CrossRef]

3.3. Multicomponent Sepsis Interventions

3.3.1. Overview

Identifying sepsis as quickly as possible is of critical importance to improving outcomes, but there are other areas of sepsis care and management that can improve outcomes, such as test ordering and results delivery, and initiation of treatment following a sepsis diagnosis. In response to this complexity, some institutions have implemented multicomponent quality improvement (QI) programs aimed at improving the full spectrum of sepsis recognition and care. Several studies found in the search results for the PSPs Patient Monitoring Systems and Screening Tools concern such multifaceted QI initiatives. We did not include these studies in the PSPs above, because it is impossible to know which elements of an initiative are responsible for any process or outcome effects. However, five such studies are briefly discussed here.

All five studies were in the hospital setting, three of them in the ED.15 All five included a manual screening tool or a PMS accompanied by an education program for clinicians and other components that varied by study. Four of the five included a sepsis-specific EHR order set so that clinicians could efficiently order the initial workup and goal-directed therapy (i.e., broad-spectrum antibiotics, IV fluid) specified in the SSC bundle. Several programs aimed to improve time from antibiotic ordering to initiation of treatment and used strategies such as ensuring that antibiotics are well stocked on the unit. One study increased the number of nurses in the ED and provided more space for triage. All studies were observational in design and therefore more prone to bias than randomized or quasi-experimental studies.

3.3.2. Evidence Summary

All five multicomponent studies reported an improvement in at least one process measure, including time to antibiotic administration or compliance with the SSC bundle. For example, Judd and colleagues found that time to antibiotic administration fell from 154 minutes to 57 minutes (p=<0.001) after implementing a screening and fast antibiotics program in all units of a 433-bed tertiary care medical center.3 Gatewood and colleagues implemented a manual screening tool, EHR alerts, and an order set in the ED of a 450-bed academic hospital, and found that SSC bundle compliance increased from 28 percent to 71 percent (p=<0.001).2

Despite these process improvements, only two of the five studies found a significant effect on outcome measures. Judd et al., described above, reported a significant reduction in ICU LOS (5.85 vs. 4.21 days, p=0.003).3 MacRedmond and colleagues reported a decrease in hospital mortality rate (51.4% vs. 27.0%, p=0.02) after implementation of a screening and order set QI program in the ED of a 500-bed tertiary care teaching hospital.4 Three studies reported absolute improvements in mortality or hospital LOS that did not reach statistical significance. One study reported an improvement in a sepsis-related mortality index, but did not report a p score or confidence interval to assess significance.1

3.3.3. Implementation

Many of the barriers and facilitators to the implementation of a multicomponent intervention are similar to those for implementing a screening tool or PMS, including the importance of clinician education to identify signs of sepsis onset and consistent protocols across hospital units. Additional facilitators mentioned in these five studies included strong teamwork among providers, pharmacy staff, and nursing personnel, and empowering the pharmacy staff to take a more active role in prescribing and ensuring initiation of antibiotics. One study found that additional nursing staff and space for triage were needed to overcome delays in diagnosis and treatment of sepsis.5

3.3.4. Gaps and Future Directions

While implementing complex QI for sepsis care is difficult to study in an evidence-based systematic review, the complexity of sepsis detection and treatment may require a multicomponent approach to reduce mortality and improve other process and outcome measures. More studies with consistent sepsis QI components and rigorous designs (randomized, quasi-experimental, etc.) would be needed to be able to review the consistent effects across studies.

References for Section 3.3

1.
Beardsley JR, Jones CM, Williamson J, Chou J, Currie-Coyoy M, Jackson T. Pharmacist involvement in a multidisciplinary initiative to reduce sepsis-related mortality. Am J Health Syst Pharm. 2016;73(3):143–9. Doi: 10.2146/ajhp150186. [PubMed: 26796908] [CrossRef]
2.
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. doi: 10.1136/bmjqs-2014-003552. [PubMed: 26251506] [CrossRef]
3.
Judd WR, Stephens DM, Kennedy CA. Clinical and economic impact of a quality improvement initiative to enhance early recognition and treatment of sepsis. Ann Pharmacother. 2014;48(10):1269–75. Doi: 10.1177/1060028014541792. [PubMed: 24982314] [CrossRef]
4.
MacRedmond R, Hollohan K, Stenstrom R, Nebre R, Jaswal D, Dodek P. Introduction of a comprehensive management protocol for severe sepsis is associated with sustained improvements in timeliness of care and survival. Qual Saf Health Care. 2010;19(5):e46–e. doi: 10.1136/qshc.2009.033407. [PubMed: 20671074] [CrossRef]
5.
Mittal Y, Sankar J, Dhochak N, Gupta S, Lodha R, Kabra SK. Decreasing the time to administration of first dose of antibiotics in children with severe sepsis. J Healthc Qual. 2018. doi: 10.1097/JHQ.0000000000000141. [PubMed: 30247207] [CrossRef]

Conclusion and Comment

The two PSPs reviewed in this chapter aim to reduce the time to recognition of sepsis so that treatment can be initiated quickly, with improvement in important patient outcomes. The review of evidence shows that manual screening tools can improve time to treatment, but the effect on mortality and other outcome measures is uncertain. Such tools may be most useful in non-hospital settings such as EMS and nursing homes, but many more studies are needed to test their effects in these settings. Evidence for PMSs in the hospital setting showed some improvement in both process and outcome measures, especially in non-ICU units. However, many studies were observational in design, limiting their strength and increasing the risk of bias. More rigorous studies are needed to test the effects of these systems.

Implementing a screening tool or PMS for sepsis requires dedicated resources and effective staff training, and it can be costly. Either type of tool can be effective if it demonstrates acceptable and sustained sensitivity and specificity, which requires pre-validation and regular monitoring. A manual screening tool is more time intensive for clinicians, but an electronic PMS may be more costly to implement and more difficult for staff to use. The customizability of a PMS’s features (e.g., “snooze” button) can add flexibility to the complexities of sepsis care, but this comes with a higher cost to implement than a manual screening tool. The decision to implement a sepsis recognition PSP, and whether it should be manual or automated, should be based on the needs and constraints of the particular setting rather than a “one-size-fits-all” approach.

Footnotes

a

SIRS criteria include: temperature higher than 100.4°F or lower than 96.8°F, HR higher than 90 beats/min, RR higher than 20 breaths/min or arterial carbon dioxide tension lower than 32 mm Hg, and white blood cell count higher than 12,000/µL or lower than 4000/µL or with 10 percent immature (band) forms.

b

Studies conducted outside of the ICU measured subsequent ICU LOS in patients who were transferred to the ICU from their unit.

Reviewers: Aline Holmes, R.N., D.N.P., Kristen Miller, Dr.P. H., C.P.P.S, and Sam Watson, M.H.A.

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (14M)

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...