Use of the quick sequential organ failure assessment tool for sepsis screening Submitted by Rodney James Buchanan A Direct Practice Improvement Project Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Nursing Practice PennWest University Edinboro, Pennsylvania December 5, 2022 © by Rodney James Buchanan, 2022 All rights reserved. PENNWEST UNIVERSITY Use of the quick sequential organ failure assessment tool for sepsis screening by Insert Rodney James Buchanan has been approved December 5, 2022 APPROVED: Robin Bilan, DNP, CRNP, RN, Project Chairperson ______________________________________________ Megan Larson, DNP, CRNP, Project Committee Member ______________________________________________ Colleen Bessetti-Barrett, DNP, CRNP, FNP-BC Project Committee Member ______________________________________________ _________________________________________ Date Abstract Sepsis is a clinical condition that requires early identification and initiation of evidencebased interventions to improve mortality and outcomes. A quality concern that has been identified is that the emergency department at the practice site is not actively screening for sepsis patients. This has resulted in difficulty meeting three-and six-hour sepsis bundle metrics. The purpose of this quality improvement project is to implement a sepsis screening methodology to correct this gap in practice. The door to lactic acid collection metric will be used to determine if any statistical improvement was made after the implementation of a sepsis screening tool. The screening tool selected for the project will be the qSOFA tool as this is one of the easiest tools to use and implement. The quality improvement project used a quantitative methodology with quasi-experimental design. The results demonstrated a statistically significant improvement in the collection times of lactic acid post-implementation. The results further support the use of sepsis screening in the emergency department setting for rapid identification and treatment of sepsis conditions. Emergency nursing plays a pivotal role in the identification and treatment of this condition as nurses are often the first healthcare professional a septic patient will encounter. Keywords: qSOFA, lactic acid, sepsis screening Table of Contents Chapter 1 Introduction to the Project ...................................................................................1 Background of the Project .............................................................................................2 Problem Statement .........................................................................................................3 Purpose of the Project ....................................................................................................5 Clinical Question ...........................................................................................................5 Advancing Scientific Knowledge ..................................................................................7 Significance of the Project .............................................................................................8 Rationale for Methodology ............................................................................................9 Nature of the Project Design ........................................................................................10 Definition of Terms......................................................................................................10 Assumptions, Limitations, Delimitations ....................................................................11 Summary and Organization of the Remainder of the Project ......................................12 Chapter 2: Literature Review .............................................................................................13 Theme one……………………………………………………………………………15 Theme two………………………………………………………………………..….20 Theme three………………………………………………………………………….34 Chapter 3: Methodology ....................................................................................................44 Statement of the Problem .............................................................................................44 Clinical Question .........................................................................................................44 Project Methodology....................................................................................................45 Project Design ..............................................................................................................46 Population and Sample Selection.................................................................................46 Instrumentation or Sources of Data .............................................................................47 Validity ........................................................................................................................48 Reliability.....................................................................................................................48 Data Collection Procedures..........................................................................................49 Data Analysis Procedures ............................................................................................49 Potential Bias and Mitigation.......................................................................................50 Ethical Considerations .................................................................................................50 Limitations ...................................................................................................................51 Summary ......................................................................................................................52 Chapter 4: Data Analysis and Results ................................................................................53 Descriptive Data...........................................................................................................53 Comparison of Minutes to Lactate Pre and Post qSOFA Implementation ..................55 Results ..........................................................................................................................57 Chapter 5: Summary, Conclusions, and Recommendations ..............................................59 Summary of the Project ...............................................................................................59 Summary of Findings and Conclusion.........................................................................60 Implications..................................................................................................................61 Theoretical Implications ...............................................................................................62 Recommendations ........................................................................................................62 References ....................................................................................................................64 Appendix A ..................................................................................................................72 Appendix B ..................................................................................................................73 Appendix C ..................................................................................................................74 Appendix D…………………………………………………………………………75 List of Tables Table 1. Comparison of Gender Pre- and Post-Intervention…………………………….54 Table 2. Comparison of Age Pre- and Post-Intervention………………………………..55 Table 3. Normality Test (Shapiro-Wilk) All Cases……………………………………..56 Table 4. Normality Test (Shapiro-Wilk) Outlier Removed Dataset………………….…57 Table 5. Homogeneity of Variances Test (Levene’s)…………………………………...57 Table 6. Independent Samples t-Test…………………………….………………….….58 1 Chapter 1 Introduction to the Project The purpose of this project is to improve the identification of potential sepsis patients at the author’s practice site in the emergency department (E.D.) setting. Early sepsis identification is a critical part in the management of sepsis as none of the recommended sepsis measures can begin if the patient has not been identified. This is not a new concept as emergency departments have been using early identification methodologies to identify ST segment elevated myocardial infarction (STEMI) and stroke patients for years to begin early life saving or disability reducing measures. However, the adoption of early sepsis recognition has been slow and therefore the management of this patient population has resulted in high mortality rates and disability post-acute phase. Furthermore, much research has been done to show that the sooner patients receive intravenous (IV) fluids and antibiotics, the lower the mortality. As this knowledge has become so common, the Centers for Medicare and Medicaid Services (CMS) has established sepsis guidelines that require hospitals to attain. These metrics are divided into two areas, three-hour bundles (Sepsis-1) and six-hour bundles (Sepsis-2). The threehour bundle focuses on the initial care of the patient and includes door to blood culture collection, door to lactic acid collection, door to first-dose antibiotic administration and door to correct IV fluid administration. Understanding these metrics all begin at the point of registration; it becomes easy to understand why a sepsis screening tool is so important. Unlike STEMI or stroke patients where an electrocardiogram (EKG) or computerized tomography (CT) scan can be used to make a quick and early diagnosis, sepsis presentations do not have a traditional presentation all the time and the source of the infection can be from many different sites such as bladder, lung, gastrointestinal (GI) 2 or post-surgical. Therefore, a physical exam in conjunction with early lab studies are needed for the rapid identification of a potential sepsis patient. The practice site currently does not use a sepsis screening tool, which has resulted in difficulty meeting all the CMS three-hour bundle sepsis metrics. Implementation of an evidence-based screening tool such as the quick sequential organ failure assessment (qSOFA) in the triage phase of care will assist in the early identification of potential sepsis patients arriving at the practice site’s emergency department. Permission to use the qSOFA tool was obtained from Dr. Seymour (Appendix A). Background of the Project The background from this problem stems from the knowledge gained that the septic disease process itself has resulted in a steady increase in volume of septic patients to the emergency department year over year (Moore et al., 2019). Additionally, delays in first-dose antibiotics have resulted in increased mortality rates (Mitzkewich, 2018). The steady increase in sepsis presentation and known mortality rates simply forces hospitals and regulatory bodies to track sepsis data and strive to perform better. The practice site only uses a standard triage screening tool called the Emergency Severity Index (ESI) to triage patients and rank them by acuity, so that providers know which patients to select first. This is a common triage scoring system used in the United States. Kwak et al. (2018) completed a study and demonstrated the use of ESI alone is not sufficient to adequately screen for sepsis and supplemented the use of the ESI score with the qSOFA screening tool and found the tool improved mortality prediction by 5.7% for ESI I patients and 8.6% with ESI II patients. These are the most critical ESI levels, but it is not uncommon to have multiple ESI II patients at once, so having an additional differentiator is important to call out the specific septic patient. It is important to note that 3 a proper ESI II score can also be applied to a cardiac or stroke patient too, which is why having a sepsis screening process in place is needed. This is an important gap to identify as the use of ESI alone does not work. Finally, Husabo et al. (2020) completed a study of Norwegian emergency departments and found the emergency departments not using a sepsis screening tool had markedly worse performance with sepsis metrics than other emergency departments that were using a screening tool. Another significant gap noted is general knowledge deficit about sepsis, treatments, and processes. Dewalder and Hulton (2020) identified that sepsis education is not required by most hospitals or licensing boards. Again, much knowledge has been required about cardiac emergencies or neurologic emergencies, which has been shown to reduce the mortality in these clinical disease processes. However, given that sepsis mortality rates are hovering around 50% demonstrates that significant gaps in knowledge that exist for this disease process (Dewalder & Hulton, 2020). Problem Statement The problem of no sepsis screening at the practice site is clear when assessing the data. Therefore, a well written problem statement is needed to help guide the project, which is explained now. It was not known if or to what degree the implementation of quick sequential organ failure sepsis screening tool would impact Centers for Medicare & Medicaid services Sepsis-1 door to lactic acid collection data when compared to current practice among adult emergency department patients arriving through triage or by ambulance. The project will serve to solve the sepsis screening gap that exists at the practice site. Current overall compliance rates with Sepsis-1 measures are 47% at the practice site and the identified root cause of this problem is no sepsis screening methodology. 4 Meyer et al. (2018) had completed a study comparing sepsis patients from 2010 against sepsis patients in 2015 and found that sepsis related mortality rates decreased from 24.1% in 2010 to 14.8% in 2015 as more hospitals were adopting sepsis screening tools as a standard in the emergency department setting. Seymour et al. (2016) completed the original study to develop the qSOFA tool based on issues with the prior commonly used tool named systemic inflammatory response syndrome (SIRS) tool. The SIRS tool is highly sensitive and therefore screened many patients’ positive that were not. This led to overwhelming emergency departments with multiple patients undergoing expensive work-ups and not having sepsis. The practice site also used to use the SIRS tool and experienced much of the same issues, which is why they stopped using the tool, but never implemented a new methodology. Seymour et al. (2016) developed an easy-to-use tool that measures only three items, which are mental status, systolic blood pressure and respiration rate. A score of two or more is a positive score that should prompt an evaluation by a provider. The reason this tool was selected from other tools in the literature is because it was the recommended tool to use by the Third International Sepsis Taskforce (Seymour et al., 2016). Additionally, the tool is simple to use by collecting information that is already gathered in the triage phase of care and ease of its use is a critical factor with compliance (Singer et al., 2017). Other tools that may perform better as it relates to sensitivity or specificity are also more difficult to calculate or require electronic health record (EHR) reform making them cost prohibitive for this project. 5 Purpose of the Project The purpose of this project is to improve the quality of sepsis care being delivered at the practice site. A strong purpose statement is needed in order to capture and frame the project, which is reviewed next. The purpose of this quantitative, quasiexperimental project is to determine if or to what degree the implementation of the University of Pittsburgh quick sequential organ failure sepsis screening tool would impact the targeted Centers for Medicare & Medicaid Services Sepsis-1 metric of door to lactic acid collection when compared to current practice among adult patients in an urban emergency department in Pennsylvania over four weeks. As stated, prior, the compliance rates with the Sepsis-1 bundle is only performing at 47%. This project assisted the practice site to return to using a screening methodology to improve patient outcomes. This project was accomplished by using a quantitative method and a quasiexperimental design. The dependent variable will be the lactic acid collection time and the independent variable will be the qSOFA tool. It is expected that by implementing the qSOFA screening tool in the emergency department, that a reduction in the time needed to collect a lactic acid level will be recognized. This is an important concept as lactic acid is one of the needed data points for a provider to diagnosis a sepsis condition. The sooner this value can be obtained, the sooner a provider can take their physical exam and add context to what they are seeing to make a diagnosis and begin treatment. Clinical Question The clinical question for this project is to what degree does the implementation of The University of Pittsburgh quick sequential organ failure sepsis screening tool decrease the lactic acid collection time when compared to current practice among adult patients in an urban emergency department in Pennsylvania over four-weeks? This question narrows 6 the focus of the project to evaluating the effectiveness of the qSOFA tool as it relates to timely collection of lactic acid in the emergency department. The variables that will be included as part of this project include an independent variable, which is the qSOFA tool and a dependent variable, which is the door to lactic acid collection time. The reason for selecting the dependent variable is related to the importance of lactic acid values in the diagnosis of sepsis (Ferando et al., 2018). By improving the time to obtaining the lactic acid level, it allows the treating provider to make decisions about completing the other Sepsis-1 metrics to improve compliance. Using this clinical question allows the team in the emergency department to understand their focus and understand the importance of timely collection of the lactic acid level. The clinical question also defines the population being used and both the independent and dependent variables that are being measured. Lastly, the clinical question sets a defined measurement period for the project. Variable Variable Type qSOFA screening tool Independent Level of Measurement Nominal Door to lactic acid minutes Dependent Ratio Quantitative (Clinical question and Variables) Q1: To what degree does the implementation of The University of Pittsburgh quick sequential organ failure sepsis screening tool impact Centers for Medicare 7 & Medicaid Services Sepsis-1 door to lactic acid draw collection when compared to current practice among adult patients arriving through triage or ambulance in an urban emergency department in Pennsylvania over four-weeks? Variable 1: qSOFA screening tool is the independent variable that is a scoring system identifying patients as either positive or negative making the scoring a nominal variable. Variable 2: Door to lactic acid measurement is a time measurement starting at zero minutes and working its way up. Having an absolute zero point makes this a ratio variable. Advancing Scientific Knowledge The project is focused on improving care of the sepsis patient in the emergency department setting. At the same time, the project will also provide advancement in scientific knowledge by implementing evidence into practice and measuring outcomes to add to the body of evidence that already exist related to using qSOFA sepsis screening in a real-world setting. This is an important concept as the translation of evidence into practice is the final step to advancing the scientific knowledge regarding a particular topic or intervention. In order to advance scientific knowledge, it is needed to have a theoretical base to build upon. This project will use Imogene King’s systems theory. King (1968) developed a grand theory called systems theory. This theory was founded on the principles of human interaction with their environment. This project recognizes the current environment does not subscribe to any type of sepsis screening methodology. This allows the nursing and physician staff to work within this environment without feeling the need for screening. However, by changing the environment and adding the expectation of 8 sepsis screening, this changes how the nursing and physician staff respond. Overtime, sepsis screening will become part of the environment and much like screening for a STEMI or stroke is rote behavior, so too will sepsis screening. As previously reviewed, the sepsis population continues to grow as our average age increases and as persons over the age of 65 become greater. Septic conditions secondary to pneumonia and pressure ulcers are common in the elderly. This means at the micro-level; every practice site needs to be focused on sepsis identification and treatment in the emergency department. The proper implementation of qSOFA sepsis screening process will at a minimum set the standard for using a sepsis screening methodology in an emergency department that in their current practice does not screen for. Solving for this gap in knowledge at the practice site will further advance the scientific knowledge within the local emergency department setting. Significance of the Project The project has great clinical significance as it is centered on the current clinical problem of sepsis and it focuses on a real gap in practice. The implications of this project are many as this could be the beginning of a culture change within the local emergency department setting by actively screening for and identifying potential sepsis patients. Evidence has shown that early identification has led to early treatment and decreased mortality. This means, the local practice site could see an increase in early sepsis identification leading to early treatment and reduced intensive care unit (ICU) admissions, shorter length of stay (LOS) and the ability to further the sepsis screening process by completing EHR reform to include automated sepsis screening technology. This project also falls in line with current evidence-based standards, which rely on early identification of sepsis patients to reduce mortality. This is an important concept to 9 note as it keeps the project in line with current evidence. By completing this project, it adds to the body of evidence not only regarding the use of the qSOFA tool, but also with the importance of sepsis screening in general. Finally, this project allowed for the incorporation of front-line and newer nursing staff with the implementation and show them the importance of completing a quality improvement project needed for sustainable change. By moving through this implementation process, this fostered the interest of and knowledge for improving practice at the bedside. Fostering this professional desire will only yield benefit for other projects or gaps in knowledge that are discovered. Rationale for Methodology The methodology for this project is quantitative. Quantitative methodology is best when implementing an independent variable and measuring the outcome (Melnyk & Fineout-Overholt, 2019). For the purposes of the project, the independent variable is the qSOFA screening tool and it will be implemented to measure impact on lactic acid collection times (dependent variable). This project is not seeking the opinion of the staff for ease of use or the qSOFA tool or their feelings about sepsis screening, which would be a qualitative methodology. The clinical question that was reviewed prior also leads to the need for a quantitative methodology. This is because the clinical question is asking if the implementation of the tool will demonstrate a measurable outcome. By virtue of the clinical question, one must use the quantitative methodology. Finally, quantitative methodology is appropriate when answering questions such as what is the problem, frequency of the problem or will this intervention improve the problem (Melnyk & Fineout-Overholt, 2019)? This project answers these questions in that the problem is poor 10 compliance with sepsis metrics and a literature review show using a screening intervention helps to improve compliance. Nature of the Project Design The design of this project is a quasi-experimental design. In an experimental design, the dependent variables are randomly separated and only the control group would receive the intervention. However, it would not be ethical to withhold an evidence-based intervention from some patients while allowing others to benefit from being screened. Therefore, a quasi-experimental design will be used to allow for the intervention to be applied to all patients and measure the outcome using all patients. The quasi-experimental design is also a good design to use for a quality improvement project. This is because a quality improvement project is not developing original research. It is translating existing research into practice and measuring the outcomes to add to the body of literature. Again, when following the clinical question, it calls for application of qSOFA screening to all patients, which further supports the quasiexperimental design. Definition of Terms The Definition of terms will be explained in this section. qSOFA Quick sequential organ failure assessment is a sepsis screening tool developed to screen for potential sepsis patients using three commonly used assessments to include mental status, blood pressure and respiration rate (Seymour et al., 2016). Clinical Significance. Clinical significance (also known as clinical relevance) indicates whether the results of a project are meaningful or not for when applied to patient care. 11 Statistical Significance Statistical significance indicates if the intervention demonstrated a measurable improvement by using descriptive statistics. This also means that the measurable improvement did not happen by chance by setting a p-value of <0.05. Intervention Group. The intervention group is the adult population that will receive the screening tool in the emergency department. As this is a quasi-experimental design, there is no control group. Assumptions, Limitations, Delimitations The assumptions of this project include a pre-existing knowledge of how to obtain vital signs and an understanding of their meaning. It is assumed that when a patient is screened positive, this triggers an action to begin obtaining things like blood cultures, lactic acid as well as other lab values needed to make a diagnosis. Additionally, it is assumed that actions such as IV fluid administration and appropriate radiological studies will also be obtained. Limitations of the study include training that may not have occurred with visiting staff in the emergency department, conflicting clinical presentations such as a stroke presentation with a positive qSOFA screening. The stroke workup would take precedence over the sepsis work up. An additional limitation was human error by not completing the paper qSOFA screen as this is not automatically incorporated into the medical record. The delimitations of the project include using the emergency department as the location for the project in order to have the greatest opportunity to use the screening tool and using the adult population as this is the most readily available population to use. Another delimitation is using a paper tool over an electronic tool since the emergency 12 department involved is part of a network of eight emergency departments and the use of the EHR would require changes to all the documentation for all the emergency departments even though they are not in scope. Summary and Organization of the Remainder of the Project To summarize, the reason for project was a gap in clinical practice was found that has resulted in the practice site not being able to meet some of their sepsis goals. After a literature review was completed, it was noted that evidence shows sepsis screening is the best way to improve sepsis metrics. A sepsis screening tool called the qSOFA was selected as this is the easiest and most cost-effective tool to implement to meet the needs of the practice site. A theoretical framework was also selected to help guide the development of the project and a change model was selected to help guide the change in the emergency department. Next a detailed review of the literature will be presented specific to the qSOFA tool and the relationship it has with lactic acid. This will show the connection between why lactic acid was chosen to be the metric to measure. 13 Chapter 2: Literature Review The purpose of this quantitative quasi-experimental project is to determine if or to what degree the implementation of the Quick Sequential Organ Failure Assessment would impact sepsis screening when compared to current practice of no sepsis screening among adult patients arriving through triage or ambulance in an urban E.D. in Northwest Pennsylvania over four-weeks. CMS has developed three-hour bundle metrics that each hospital in the United States is to work to attain. These metrics include door to lactic acid collection, door to blood culture collection, door to first dose antibiotic administration, door to 30 ml/kg IV fluid bolus. The first step in being able to meet these metrics is identification of a potential sepsis patient presenting to the E.D. The international consensus sepsis task force is the international organization that researches sepsis and meets to update guidelines every four-years. During the meeting in 2016, the task force made a recommendation to begin using the qSOFA sepsis screening tool instead of the SIRS tool that was recommended in the 2012 meeting. Seymour et al. (2016) completed the original study that developed the qSOFA tool and found that qSOFA had a much higher specificity for predicting 30-day mortality than SIRS, however qSOFA was noted to be less sensitive than SIRS. These findings led the sepsis task force to make the recommendation to transition from a high sensitivity tool to a high specificity tool. Theoretical Foundations This project seeks to correct a systems issue. Therefore, Imogene King’s system’s theory will serve as the grand theory to guide the project. King (1968) explains that nursing is a process of action, reaction, interaction, and transaction by which care to humans is delivered. These actions all occur within a social institution resulting in three distinct levels of operation which are, the individual, the group and society (King, 1968). This 14 theory fits nicely within this project as the nursing staff are the group using the tool. The individual nurse functioning as an individual will make personal choices on many things from how they see themselves as part of the solution, commitment to learning the functioning of the tool and accurate application of the tool. However, the nurse is not alone in this project, they will also be functioning as a group. Nurses will be using the tool, but also relying on other nurses for support, validation, and mastery of competence. Finally, the nurse will be functioning within the societal structure of the hospital that currently does not use a screening methodology for sepsis. These levels of operation will be occurring fluidly and in conjunction with each other as a system. Change is also an important aspect to consider when implementing this project. Kotter (1996) developed an eight-step change model used in leadership to implement change. The first step is to establish urgency by identifying realities and acknowledging major opportunities, second is creating a guiding coalition by establishing a group with the authority to make change and lead a team, third is developing a vision so the team understands the end goal and can develop strategy to implement change, fourth is communication of the vision to the department, fifth is empowering broad-based action by eliminating obstacles and encouraging out of the box thinking, sixth is generation of short term wins to fuel motivation, seventh is consolidating gains from lessons learned and capitalizing on results and finally the last step is anchoring the new approach by linking the new outcomes/results with the new behaviors (Kotter, 1996). Review of Literature A literature review was conducted to identify the problem and establish themes. Google Scholar, Pubmed and TRIP database were used with the following search terms “qSOFA,” “qSOFA screening”, “sepsis screening ED”, “sepsis mortality reduction”, 15 “qSOFA and lactic acid” and “sepsis alert”. The publication range was set 2017-2021 and full text methodology was used to find articles. The search yielded 370 articles, but only 40 were selected. Of the 40, 10 were determined to be quasi-experimental studies and were discarded from the literature review. While reviewing the 30 articles several themes began to emerge beginning with the importance of lactic acid and what it tells clinical providers when caring for a sepsis patient. The second theme was the various amounts of screening tools to use. The final theme was the low sensitivity of the qSOFA screening tool compared to many of the other tools that have been developed. Theme 1, Lactic Acid The first emerging theme was the importance of lactic acid. As stated earlier, this is one of the CMS metrics in the three-hour bundle (door to lactic acid collection). Gattinoni et al. (2019) completed a study to determine what causes the elevation of lactic acid in sepsis, which they found to be impaired tissue oxygen use, instead of impaired oxygen transport. This is an important concept in the management of sepsis, because this pathophysiology can be reversed with IV fluid resuscitation, but only if sepsis is recognized and lactic acid is measured. For these reasons, the learner’s project will be focused on the door to lactic acid measurement time metric. Subtheme 1, Lactic Acid Levels Understanding the importance of lactic acid values. This subtheme emerged from literature regarding the understanding about various levels of lactic acid. This understanding leads to a deeper knowledge base regarding sepsis. Fernando et al. (2018) completed a study to determine if lactic acid levels of 2-3.9 16 in septic patients would result in a lower level of acuity and allow for placement on a lower acuity floor instead of an ICU. The study was a prospective observational cohort design that enrolled patients over the age of 21 who met the 1992 society of critical care medicine (SCCM) definition for sepsis. The sample size was 985. Criteria set up for inclusion included meeting SCCM criteria, lactic acid level of 3.9 or less and not receiving any critical care interventions in the E.D. All clinical data was reported using 95% confidence intervals and area under receiver curves (AUC) were used to determine the predictive power of the lactic acid level. Limitations of the study included the decision to use critical care definitions from 1992. Another limitation of the study was that approximately 20% of the enrolled patients did not have a lactic acid level drawn due to physician choice. The results showed that patients with lactic acid levels of 2-3.9 had a specificity rate of 66% for deterioration within 72 hours of arrival. These results show that even at lower levels of lactic acid in sepsis patients, the risk of clinical deterioration remains high, which is why obtaining a timely lactic acid is so important for this population. Recommendations for future research would be to complete a study determining lactic acid clearance by following serial lactic acid levels. As stated earlier, Gattinoni et al. (2019) completed a study to better understand and explain how elevated lactic acid performs from a pathophysiological perspective. This was a post hoc analysis of a random control trial completed analyzing the impact of using albumin for fluid resuscitation in sepsis. Taking the patient sample of 1741 ICU patients, the authors already had lactic acid and central venous oxygen saturation (Scvo2) information. Using this 17 information, the base excess was calculated to better understand the relationship between hyperlactatemia and acidemia. The authors used AUC were used compare continuous variables and chi-square test were used to compare differences between groups. A limitation of the study was since the authors had used the sample from a different study, they did not know the volume of IV fluids that had been administered in the E.D. setting, which could impact the lactic acid levels of the population, which was an ICU population only. The results showed that elevated lactic acid levels in sepsis were mainly caused by impaired tissue oxygen use instead of impaired oxygen transport. This demonstrates that sepsis causes impairment at the tissue level which is an important concept to understand in the management of sepsis. This also validates why CMS has placed the door to lactic acid metric in the three-hour bundle. Subtheme 2, Enhancing Quality of Screening Tools The second subtheme determined from the lactic acid theme is the important role lactic acid plays in enhancing the quality of screening tools. This is a novel concept that has been emerging in the literature. This concept stems from having a better understanding of the pathophysiology of elevated lactic acid levels. Baumann et al. (2020) completed a study assessing positive qSOFA scores and elevated lactic acid levels. The authors used a retrospective study design with a sample size of 3743. The data was then sorted by lactic acid level of greater than two and greater than four. The authors used AUC for qSOFA positive patients with both separated levels of lactic acid. The results showed that qSOFA scores greater than 1 and a lactic acid level of greater than 2 showed a sensitivity of 94%. This is much higher than using qSOFA alone. The sensitivity rate of qSOFA 18 score three and lactic acid level greater than four was 7.1%, which makes sense as a patient would have to attain all qSOFA points and have a very high lactic acid in order to be positive. Limitations identified included missing data due to the retrospective design, patients discharged from the E.D. were excluded from the study, which means the combination of qSOFA and lactic acid could not be used on all patients arriving to the E.D. The authors suggest that additional research needs to occur to establish a relationship between lactic acid and qSOFA. Jung et al. (2018) completed a study to determine if changing the qSOFA scoring tool by adding an additional point for lactic acid level would improve the sensitivity of the tool. This was a retrospective study with a sample size of 1226 patients that required emergency abdominal surgery for a complicated intraabdominal infection. These records were used to determine qSOFA and SOFA scores. The points of measurement for qSOFA are respiration rate greater than 22, systolic blood pressure less than 100 and altered mental status defined as a Glasgow Coma Score (GCS) of less than 15. The patient receives one point for each of these measures and a score of 2 or more is considered positive. The authors gathered lactic acid measurements for each of these patients and made an additional point if the lactic acid level was greater than two. The t-test and MannWhitney test were used to measure continuous variables. The chi-square test was used to establish relationship between categorical variables. To determine the sensitivity, the AUC was used. Limitations of the study included a surgical population presenting the E.D. that typically are ranked higher for acuity and often do not follow a standard protocol and this was a retrospective study and the results may not be able to be generalized to all hospitals. The results showed the 19 qSOFA sensitivity rate was 46% in this patient population. The qSOFA plus lactic acid tool had a sensitivity rate of 72%, which aligned with other like studies. The authors suggest that prospective studies of patients presenting to the E.D. with suspected abdominal infection should be performed to gain further insight on modifying the qSOFA tool. Subtheme 3, Latic Acid and Clinical Decision Making The third subtheme determined from the lactic acid theme is the importance lactic acid levels play in clinical decision making. This is an important subtheme as lactic acid levels assist in the decision making for providers on how aggressive they should treat and placement of the patient after the E.D. episode of care. Aksu, et al. (2018) completed a study by taking the qSOFA screening tool, which has three measured items which are, altered mental status, elevated respiration rate and low blood pressure and they added a “fourth” point being a lactic acid level of 2.3. The results showed by adding the additional lactic acid point, this improved the qSOFA tool’s ability to predict 30-day mortality by 50% and enhanced the clinical decision making for the treating provider to treat more aggressive and more accurate placement of a patient for inpatient admission. This was an observational, prospective study that used a population of adult patients 65 years and older that were admitted to the intensive care unit (ICU). Their sample size was 799. The authors then determined lactic acid levels that would function as another point on the scoring, which were 2.4 and 4.0. The authors used MannWhitney tests to compare the groups and chi-square test to summarize the categorical variables. Then to perform weighing of the new score, they used a multiple regression analysis for the outcome of mortality. Finally, AUC was used 20 to determine the statistical significance. The results showed that when the lactic acid level of 2.3 was added to the score for both the qSOFA and SOFA score, both tools demonstrated a statistically significant higher predictive value to predict mortality. Specific to the qSOFA tool, a lactic acid level of at least 2.3 and a score of 2 resulted in a 50% mortality rate. Understanding these results allows providers to treat these patients with urgency, much like a stroke or trauma patient. Limitations identified was that this was a single site study and the results could not be generalized to all facilities and this facility also saw a higher-thanaverage amount of cardiac patients than other facilities around them, which could have skewed data due to the higher than usual groin access procedures. The authors recommend additional research on this topic as this would be a change to the design of the qSOFA tool. Shetty et al. (2018) completed a study is to determine at what point the lactic acid level should be considered positive when a patient is presenting with possible sepsis to an E.D. This was a retrospective study using the Australian SEPSIS KILLS government program that oversees sepsis care in Australia including an electronic sepsis registry. Data was obtained from this registry for adult patients 18 years and older that were cared for in the E.D. The limitation identified in the study was that data was used from SIRS screening, which is no longer the recommended screening tool. The results revealed a mortality rate of 8.4% for a lactic acid level of greater than 2 even with systolic blood pressure greater than 100. Mortality rates increased to 19.7% with a lactic acid level greater than two and systolic blood pressure less than 100. The analysis of the data shows that lactic acid levels of two and higher should be used as a positive 21 cutoff point to consider potential sepsis especially in the presence of hypotension. This should prompt clinical providers to have an increased awareness when lactic acid levels are present. The authors suggest completing the study using qSOFA data to establish a relationship between lactic acid and qSOFA scores. Theme 2, Importance of Screening Tools The second theme that emerged from the literature review was the importance of screening tools in the recognition and treatment of sepsis. Without recognition, sepsis care cannot begin. When reviewing evidence, it becomes clear that many different tools exist and each of them have their positive and negative attributes. Seymour et al. (2016) completed the original study that produced the qSOFA tool and it is noted that this tool is easy to use and a good starting point for a facility that currently does not do sepsis screening. Therefore, this tool was selected for the learner’s project. Several subthemes have been identified and will be reviewed. Subtheme 1, qSOFA Screening Tool The first subtheme is the use of the qSOFA tool itself. This is the largest subtheme under the second theme for obvious reasons as this is the tool that has been selected for implementation. During the review of literature under this subtheme, it will be noted that many studies are comparing the qSOFA tool against other tools such as SIRS, sequential organ failure assessment (SOFA) and national early warning score (NEWS). Other studies simply review the performance of qSOFA independently. Barbara et al. (2018) completed a study to evaluate the use of qSOFA in the pre-hospital environment. This study is a retrospective study that sampled adult patients 18 years and older that were transported to one of the two hospitals 22 in the study. These patients were already identified as qSOFA positive (score of two) by the ambulance crews. The sample size was 271. These patients were then compared to their admitting diagnosis to determine if they were admitted for sepsis, severe sepsis, or septic shock. The patients were then separated into two groups, one group being sepsis and the other being non-sepsis. Differences in the groups were assessed using the t-test and comparisons of skewed data was analyzed using the Mann-Whitney test. Finally, the chi-squared test was used to compare the variables between the two groups. The largest limitation to the study was not having denominators in both groups, meaning not having patients that were septic or non-septic, but not screened pre-hospital as positive. This prevented the authors from determining sensitivity and specificity data. The results showed that 66.67% of the patients that were scored positive by prehospital services were indeed admitted with sepsis, leading the authors to suggest further study of the qSOFA tool in the pre-hospital setting. Burnham & Kollef (2018) completed a study to determine if the use of the qSOFA tool has resulted in inappropriate use of initial antibiotics. This is a retrospective study that collected data from records on adult patients hospitalized with a Enterobacteriaceae infection. qSOFA scores were determined based on the worst vital signs in the record 24 hours prior to becoming blood culture positive. All antibiotics prescribed and administered were collected. The chi-square test was used to determine categorical values and the t-test, Mann-Whitney and oneway ANOVA was used to determine continuous variables. The sample size was 510. Limitations of the study include being a single-center study leading to results that may not be generalizable and altered mental status may have only been 23 documented in the most severe cases, which could inflate mortality. The conclusion of this study reveals the qSOFA can be used as predictor of mortality and the results of the study show no impact of a positive score related to inappropriate antibiotic selection. The authors suggest that further research should occur regarding the importance of lactic acid levels in the sepsis population. Canet et al. (2018) completed a study using the same study methodology that Seymour et al. (2016) used, but only using E.D. patients to determine if the qSOFA tool can predict mortality upon presentation in the E.D. setting. Seymour et al. (2016) used an in-hospital population, to develop the qSOFA tool, so the results could not be generalizable to the E.D. population. This study is a retrospective cohort study that developed a query of the EHR to obtain all E.D. records of adult patients 18 years and older that were screened for suspected infection. The sample size was 165,912. The statistical tests used were the Wilcoxon rank sum test. Limitations of the study included this being a singlecenter study, so the results cannot be generalizable to all centers. The results of the study showed that for patients who screened positive in the E.D., the qSOFA tool had a 61% sensitivity rate and 80% specificity rate combined with a high mortality rate. The authors recommend further research on evaluating the qSOFA tool in the E.D. setting. Goulden et al. (2018) completed a study is to compare predictive mortality rates of qSOFA, SIRS and NEWS screening tools. This is a retrospective cohort study using adult patients that presented to the E.D. with suspected sepsis. The sample size was 2158. The statistical analysis used the AUC to determine sensitivity and specificity as well as positive and negative predictive value for 24 mortality. Limitation of the study was the authors did not have data on comorbidities or cause of death, so this makes it impossible to determine that patients who died, did so from sepsis specifically. The results showed that for predicting mortality qSOFA was most specific followed by NEWS then SIRS. However, SIRS was the most sensitive followed by NEWS then qSOFA. This determines that none of these tools score high for both sensitivity or specificity. The authors suggest further research be conducted comparing the various scoring tools for sepsis. Kievlan et al. (2018) completed a study is to determine if completing the qSOFA score in six-hour increments during the first 48-hours would demonstrate improved predictive validity of the tool. This was a retrospective cohort study that used the original qSOFA study population of 1,309,025 patients. These patient records were then used to develop trajectory modeling using their vital signs data to apply a qSOFA score every six-hours. The standard deviation and median were calculated for the continuous variables. To compare predictive validity the authors used the Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD) recommendations. Finally, AUC were used to show differences in scores at each measurement interval. The authors identified limitations of using only an in-hospital population and they did not compare predictive validity against other screening tools. The results showed that the repeated measurements improved the AUC curve from 0.79 to 0.85 in predictive validity to predict mortality suggesting that qSOFA should not be considered a one-time use only scoring tool. Future suggestions for research were to study the 25 repeating of the qSOFA score until discharge to determine the trajectory of qSOFA throughout the hospital stay. LeGuen et al. (2018) completed a study to determine how many rapid response patients will have a positive qSOFA score, how antibiotics were started based on the score and if positive qSOFA patients also had an infection. This was a prospective observational study that reviewed every rapid response call from June 6th, 2016 to July 10th, 2016 for adult patients over the age of 18. The sample size was 258. Statistical analysis used the chi-square test to compare variables between groups and two-sided p-value to indicate any statistical significance. The authors identified the limitations of a small sample size and the possibility of selection bias secondary to the time of the year. The results showed that qSOFA positive patients had a mortality rate of 39% compared to non-positive patients of 16% regardless if they had an infection or not. For patients that did have an infection and were scored positive their mortality rate was 26% compared to qSOFA negative at 15%. The analysis of the data demonstrated the qSOFA scoring tool was useful in assisting the rapid response team in the identification of sepsis and critical patients alike. The authors suggest further study of the qSOFA tool at other facilities for rapid response patients. Seymour et al. (2016) completed a study to develop a screening tool that had a better specificity than the currently used SIRS tool. This was because the third international consensus sepsis task force had determined the current recommendation of using the SIRS tool was yielding a very low specificity, making it an unreliable tool. This was a retrospective cohort study that collected data from 2010 to 2012 using databases from 164 hospitals with a sample size of 26 4,885,558. The authors used three different sepsis scoring methodologies SIRS, SOFA and Logistic Organ Dysfunction System (LODS) and scored the records based on established algorithm. Using these scores and measuring the datapoints that makes these scores, the authors were able to assign weights to the specific clinical criteria that were most significant to each score to predict mortality. The three things that weighed the most was elevated respiration rate, low blood pressure and altered mental status. This aligned the closest with the SOFA scoring and the authors coined the term "quick" SOFA. The statistical analysis of used was quite complex using such a large database, but the authors used AUC to measure sensitivity and specificity of the new tool. The authors identified several limitations. The first was using only patients with a diagnosis of sepsis. Secondly, the tool does not have any lab values in it to make the sensitivity higher and finally using altered mental status can be difficult with patients presenting with dementia. The results showed an AUC for qSOFA of 0.71 specificity to predict 30-day mortality in the non-ICU patients. qSOFA scores of two or three accounted for 70% of deaths and 70% of ICU stays longer than three days. The authors suggest studies of the tool need to be completed internationally to determine its real-world usability. This subtheme reveals the many types of screening tools that are available to be used. This subtheme also demonstrates that not a single tool being used is perfect, including the tool being recommended. However, as with any type of clinical scoring tool, this does not replace clinical judgement and physical exam. 27 It simply calls the patient out, so that the physical exam can occur in a timelier manner. Subtheme 2, qSOFA and Special Populations The second subtheme that emerged from the second theme was qSOFA’s performance in special patient populations. Discussions about sepsis in the E.D. setting are generally broad, because the E.D. is designed to be an all-comer department. This raises the question about how qSOFA would perform specifically with pneumonia and cancer patients. These sub-populations are of particular interest, because pneumonia can mimic congestive heart failure (CHF) or congestive obstructive pulmonary disease (COPD), neither of which is an infectious process. Cancer is also a population that will often have blunted sepsis responses secondary to chemotherapeutic treatments and steroid therapy. Asai et al. (2019) completed a study is to determine if the qSOFA tool can evaluate the severity and prognosis of patients with pneumonia against traditional pneumonia scoring systems. The population was all patients that presented to the hospital between 2014-2017 with either community-acquired or healthcare-acquired pneumonia (nursing home or rehab facilities) were entered into the study. This was a retrospective study with a sample size of 257. Each record was scored using the SOFA, qSOFA and traditional pneumonia screening tools used by the authors using admission data found in the record. The authors used a chi-square test to compare the categorical variables. The Mann-Whitney test was used to compare continuous variables and finally a logic regression was used to identify independent risk factors associated with 30-day mortality. The authors identified limitations such as small sample size from a single-facility and the population had 28 a higher amount of healthcare-acquired pneumonia versus community-acquired pneumonia. The results showed that a positive qSOFA score of greater than two correlated well with other validated pneumonia scoring tools. However, qSOFA scores of zero or one did not correlate as well. The authors felt that qSOFA should not be used solely to determine severity and prognosis of pneumonia patients, but when a qSOFA score is positive, the clinical team should also calculate a SOFA score, because their data showed a positive qSOFA and positive SOFA correlated with high levels of mortality. The authors also explain the ease of calculating a qSOFA score making it an easy tool to begin with upon arrival in an E.D. and suggest that further research using these tools to identify pneumonia should be completed. Costa et al. (2018) completed a study is to assess the qSOFA, SOFA and SIRS screening tools as it relates to mortality for cancer patients admitted with sepsis. This is a retrospective study that reviewed records for cancer patients admitted to the ICU setting with a diagnosis of potential sepsis. The sample size was 485 patients. Chi-squared test were used to compare categorical differences between groups and AUC were used to calculate results and display performance of each tool. The authors identified limitations of the study such as this being a single-center study, so results cannot be generalized to all facilities. Additionally, the authors acknowledge they were only focused on cancer patients, so comparative performance of qSOFA, SOFA and SIRS was not able to be done with other patients. The results showed that the SOFA score was more sensitive, but less specific than SIRS in its ability to predict ICU mortality. qSOFA was more sensitive, but performed equally regarding specificity compared to SIRS in 29 its ability to predict ICU mortality. There was no difference between SOFA and qSOFA to predict out of ICU mortality. The authors feel that SOFA and qSOFA scoring tools perform better than SIRS to predict mortality in cancer patients presenting with sepsis. The authors also feel that additional research needs to be completed in this population using these tools as they are only aware of one other study like theirs. Mecham et al. (2018) completed a study to examine the predictability of 30-day mortality rates of pneumonia patients using the SIRS, qSOFA and SOFA scores. This is a retrospective study that evaluated patient records of adult patients that presented to the E.D. and diagnosed with pneumonia from three 12-month time frames December 2009 - November 2010, December 2011 - November 2012 & November 2014 - October 2015. The sample size was 11,051. The authors used the AUC to determine the 30-day mortality for all three scores. t-test and Wilcoxon rank sum test were used for all time metrics and chi-square test was used to compare variables between groups. Limitations discussed in the study included missing data to calculate the SOFA score on some patients and the authors used a GCS score less than 14 to determine altered mental status, when the tool is designed to score altered mental status with a score less than 15. The results showed for 30-day mortality qSOFA had an AUC of 0.7 and SIRS 0.61 showing that qSOFA outperformed SIRS in its ability to predict mortality. Using the scores to predict admission form the E.D. qSOFA scored an AUC of 0.7 and SIRS 0.67, this translates to 42% of positive qSOFA patients being admitted and 20% of SIRS patients being admitted. These results show the qSOFA only has a moderate sensitivity, but a much higher specificity that SIRS in the recognition of 30 pneumonia. The authors suggest further research regarding changing the scoring thresholds of the qSOFA tool to improve its sensitivity. Tokioka et al. (2018) completed a study is to compare the qSOFA tool against two known pneumonia screening tools called the Confusion, Urea nitrogen, Respiratory rate, Blood pressure, age 65 (CURB-65) and the Pneumonia Severity Index (PSI) to determine the qSOFA's ability to predict ICU admission and mortality. This is a secondary study using data from a prospective study of adult patients 18 years and older that were admitted with a diagnosis of pneumonia. The sample size was 1954. Categorical variables were calculated using the chi-square and the C statistic was used to determine the probability. Limitations presented in the study included this being a single-center study, so results cannot be generalized to other centers. Additionally, at the time this article was published only four studies had been completed regarding the use of qSOFA in pneumonia patients leading to insufficient literature review by the authors. The results showed that for predicting ICU admission, positive qSOFA was 23.4%, CURB-65 was 60.8% and PSI was 59.9%. C statistics for predicting mortality for qSOFA was 0.69, CURB-65 was 0.75 and PSI was 0.74. The sensitivity of qSOFA was 39% and specificity was 88% for predicting mortality. CURB-65 and PSI both showed high sensitivity and moderate specificity, but actual percentages for these scores was not published. This means the CURB-65 and PSI scores do a better job at predicting pneumonia in the E.D. setting, but these tools are more difficult to apply to every patient coming to the E.D., therefore the use of the qSOFA was determined to be a good starting point and then apply other more specific tools during the care of the patient. The authors suggest prospective 31 observational studies need to be completed to fully understand the application of qSOFA in the pneumonia patient. This subgrouping of research articles shows that qSOFA can be used to identify potential sepsis in cancer and pneumonia patients. A common theme is that once identified, the clinician should also complete a more comprehensive scoring tool to further narrow down the predictive mortality. However, since the qSOFA is easy to use, this does offer a starting point in the E.D. setting. Subtheme 3, Importance of Using a Screening Tool The third subtheme under screening tools is the importance of using a screening tool. All the new antibiotics, ICU therapies and evidence-based interventions will mean nothing if the patient is not identified. While this sounds simple, it can be quite difficult in an E.D. setting, because all clinical presentations and age group arrive to the E.D. in an unpredictable manner requiring an E.D. to have some sort of screening methodology to detect sepsis. This is an important subtheme for the learner’s project as the practice site does not use a screening methodology to screen for sepsis. Husabo et al. (2020) completed a study is to determine if delays occur in Norwegian E.D.'s that prolong the identification of sepsis and the following treatment. This was an observational study involving 24 E.D.'s in Norway. The study used SIRS criteria to determine if a patient was positive for sepsis and this information was retrieved from a national patient registry. The sample size was 1559 patients. All data was calculated to confidence intervals of 95%. Linear regression models were used to determine variables in time metrics. A logic regression model was used to estimate the 30-day mortality rates. Limitations 32 identified were the use of the SIRS screening tool, when the recommendation is to use qSOFA. Additionally, the authors did not have data on the severity of sepsis to establish correlation between antibiotic delays and mortality. The results showed that 72.9% of patients were triaged in the first 15 minutes of arrival, 44.9% were evaluated by a provider in accordance with triage level, 83.6% had a set of vital signs obtained within the first hour. 80% of patients had regular blood work drawn in the first hour, only 48.6% of patients had a lactic acid level drawn in the first hour. 25.4% of patients received antibiotics in the first hour and 55.5% of patients received antibiotics in two hours. This poor performance was related to either no screening or ineffective processes in their E.D.’s. This study highlights the importance of leadership establishing evidence-based practices in E.D.'s to properly screen patients for potential sepsis. Meyer et al. (2018) completed a study to better understand the relationship over time of post-discharge outcomes and hospital readmission since the hospital readmissions reduction program (HRRP) has been put into place. This study is a retrospective design that used International Classification of Disease, version-9 (ICD-9) codes to pull patient records that were end-coded as sepsis, severe sepsis, and septic shock at three network hospitals. The sample size was 275,600. Limitations identified included the study hospitals were using different tools for identifying sepsis patients in their E.D.’s and the ICD-9 codes only represented the sickest and not all sepsis patients that visited the E.D. and discharged. The results showed that sepsis hospitalizations increased from 3.9% in 2010 to 9.4% in 2015. However, in-hospital mortality has decreased from 24.1% in 2010 to 14.8% in 2015. The study shows that advancements in early recognition and 33 improved clinical guidelines encouraged by the HRRP initiative have resulted in an overall decrease in mortality. The authors suggest further research at other centers to correlate their findings. Rosenqvist et al. (2020) completed a study is to measure the results of modifying the local Swedish triage scoring methodology by adding fever of 38 degrees Celsius as a trigger to initiate an evaluation by a provider to decrease the door to antibiotic metric in sepsis patients. This was an interventional study comparing outcomes before and after changes were made, but also calculating sensitivity for their tool. Patients were identified and included in the study before the change to the triage score from January 1st, 2015 to March 31st, 2015. Then a second group was studied from January 1st, 2017 to March 31st, 2017 post changes to their scoring system. The sample population for the 2015 group was 443 and the sample population for the 2017 group was 533. The authors used ttest to measure continuous variables and chi-square test to compare variables between groups. Limitations identified included adding a temperature may not pick up cancer patients or immunosuppressed patients and the authors were using sepsis definitions from 2012 despite new definitions being released in 2016. The results showed that the 2015 sample of patients received their antibiotics within one hour 68.3% of the time. The 2017 sample of patients received their antibiotics within an hour 89.3% of the time demonstrating the importance of screening. The authors suggest further research on screening tools and their criteria. Shah et al. (2018) completed a study is to measure three-hour sepsis bundle compliance before and after implementing an institution-based sepsis screening methodology and determine if this tool has any impact on 30-day 34 mortality. This was a retrospective cohort study that included all adult patients 18 years and older that were admitted with a diagnosis of severe sepsis or septic shock. The pre-implementation period was August 2012 to January 2013 and the post-implementation period was January 2015 to June 2015. The sample size for the pre-implementation group was 58 and post-implementation group was 57. Continuous variables were measured using the t-test, differences between data sets were measured using the chi-square test. A logistic regression analysis was conducted with p-values to determine mortality rates. Limitations identified included using an unvalidated screening tool and small sample size. The results showed no difference in three-hour bundle compliance for any metric except the door to first dose antibiotic metric. Pre-implementation antibiotic administration in one-hour was 10.3% and post-implementation was 89.5%. The screening tool did not show any statistical difference in 30-day mortality rate preimplementation or post-implementation. The authors suggest comparing their tool against the qSOFA tool for future research. This subtheme shows the importance of using a screening methodology. Facilities that did not use any screening noted to have unacceptable delays in treatment and antibiotics. Other facilities that did use screening tools showed superior performance. These findings in literature support the learner’s project. Theme 3, Sensitivity and Specificity of Screening Tools The final theme from the literature review is the sensitivity and specificity of the sepsis screening tools, but in particular, the qSOFA tool. Multiple studies including the original study that developed the tool have demonstrated a low-sensitivity, but highspecificity for qSOFA. SIRS screening tool was the opposite, the tool has a high- 35 sensitivity, but low-specificity (Seymour et al., 2016). This resulted in a high volume of patients screening positive thus leading to increased resource demand, but not having sepsis. Literature reviewing the low sensitivity, high specificity and ease of use will reviewed. Subtheme 1, Low Sensitivity of qSOFA Tool The first subtheme is low sensitivity of the qSOFA tool. Literature will be reviewed demonstrating this finding about qSOFA. Implications of having low sensitivity will be discussed throughout the literature review. Abdullah et al. (2020) completed a study is to evaluate the role of qSOFA as a prognostic factor for 30-day mortality in patients with suspected or proven infection fulfilling the SIRS criteria for sepsis on admission to the E.D. This study used a retrospective review of patients with sepsis diagnosis and positive SIRS criteria. These records were then reviewed and assessed to determine what the qSOFA score would be. The population size was 464 patients. Limitations identified included selection bias as the authors were using a historical database and the authors did not always have GCS data to calculate a qSOFA score. The results showed that 16.8% of the patients had a positive qSOFA score upon presentation to the E.D. Of the qSOFA positive patients, 81.8% of them needed ICU level care within the first 24 hours of their admission. The qSOFA positive patients had a 21.9% in-hospital mortality rate and a 32.9% 30-day mortality rate. A 16.8% capture rate is low for recognition of sepsis. The authors do suggest 36 completing a similar study with a larger sample size to determine if the results would be the same. Filbin et al. (2018) completed a study is to evaluate the Shock Precautions on Triage (SPoT) and qSOFA sepsis screening tools and to determine the importance that vital signs have in the early identification of sepsis in the E.D. setting. This was a retrospective observational study using ICD-9 records that were end coded for sepsis in 2009. These records were then assigned a SPoT score and qSOFA score at the time of their first set of positive vital signs. The sample size was 2669. Limitations identified included not always having the GCS score available to calculate the qSOFA score and a single-center study, so results cannot be generalized to all facilities. The results showed that the SPoT scoring yielded a 56% sensitivity rate when scored in the triage phase (initial phase) of the E.D. visit. qSOFA sensitivity was 28% in triage. The authors suggest additional research on the impact of antibiotic delays as this is a metric in the three-hour bundle, as there is not a lot of research on therapeutic benefit of this metric. Ortega et al. (2019) completed a study is to compare the prognostic value of the qSOFA tool to identify potential sepsis patients arriving in the E.D. setting as compared to using SIRS, NEWS or ESI scoring. This was a prospective allcomer study that used patients that presented to the E.D. over the age of 18 except obstetric and ophthalmologic patients. The sample size was 2523. Limitations identified include only 1.6% of the patients in the sample had sepsis and 32.3% of their sample had incomplete qSOFA scores. The results showed that 33.6% of the patients required admission to the hospital and 5.4% of those needing admitted required ICU level care. 1.5% died in the hospital and 2.5% died within 30 days. 37 The AUC for prognostic ability was as follows: qSOFA-0.79, SIRS-0.81, NEWS0.85 and ESI-0.77. These results show that qSOFA had the lowest sensitivity scores besides the emergency severity index (ESI), which is not a sepsis screening tool. The authors suggest further research to find a tool that has improved sensitivity and specificity for the E.D. setting. Perman et al. (2020) completed a study to take a known population of patients that have already been diagnosed with sepsis and have a high rate of mortality and assign a qSOFA score to them for the purposes of determining sensitivity within a known population. This was a retrospective, observational study that enrolled patients diagnosed with severe sepsis that were admitted from the E.D. The sample size was 2859. Continuous data was calculated using the ttest. Categorical variables were analyzed using the chi-square test. Sensitivity and specificity results were calculated using AUC. Limitation identified included using the worst vital signs and not the triage vital signs when the tool would be used and this was a single-center study, so the results cannot be generalizable. The results showed the qSOFA performed in triage demonstrated an AUC of 0.59, which the authors felt was not very high given these were already known sepsis patients. The authors suggest further research to develop better screening tools. This subtheme shows that the qSOFA tool has a low sensitivity, meaning patients presenting with sepsis can be missed. However, the practice site currently does not use any type of screening methodology, so even a low sensitivity tool 38 theoretically would yield results. It is also clear that additional research needs to be completed to develop a tool that has a higher sensitivity that is practical to use. Subtheme 2, High Specificity of qSOFA The second subtheme under sensitivity and specificity is the high specificity of qSOFA. In the prior subtheme, it was noted that qSOFA has a low sensitivity that can result in missing septic patients. However, qSOFA does have a high sensitivity, meaning that when a patient is screened as positive, there is a high likelihood they will be septic. This section will review literature discussing this feature of qSOFA. Brink et al. (2019) completed a study is to evaluate the 10 and 30-day mortality predictability rates for the qSOFA, SIRS and NEWS scores. This was a retrospective study that collected all E.D. records for patients presenting to the E.D. with suspected sepsis. The sample size was 75,428. Patient characteristics were compared by using a two-sampled t-test and Mann-Whitney test. A chisquared test was used to compare the distribution of the data. The AUC was used to display observed mortality. The Youden's J statistic was used to determine sensitivity, specificity and positive or negative predictive value for each test. Limitations identified included the facility where the study was completed specializes in congenital and acquired immunodeficiencies thus altering presentations and not being able to generalize the results of the study to all facilities. The results showed that the NEWS score performed best in predicting 10 and 30-day mortality, qSOFA showed the highest specificity and SIRS had the highest sensitivity. The authors suggest future research to include possibly 39 changing the scoring cut off of the qSOFA from 2 to 1 to improve sensitivity as using the NEWS tool is complex and could lead to non-compliance. Rodriguez et al. (2018) complete a study is to determined sensitivity and specificity for both the qSOFA and SIRS screening tools. This is a retrospective study that was conducted using five network connected E.D.'s on the west coast of the U.S. Patients age 17 and older that were admitted through any of these E.D.'s for suspected sepsis were identified as the study population. The sample size was 3743. Statistical analysis used AUC to calculate sensitivity and specificity. Limitations identified were the inclusion of patients that needed vasopressor support, which would favor qSOFA as a blood pressure less than 100 is one of the criteria. The results showed that a positive qSOFA had a 53.5% sensitivity rate, positive SIRS had an 86.7% sensitivity rate. For specificity, positive qSOFA had an 89.1% specificity rate, Positive SIRS had a 45.6% specificity rate. These results support the high specificity of qSOFA, which the authors deem valuable when making disposition decisions in the E.D. setting. A future research question proposed was comparing qSOFA against other scoring tools such as NEWS. Tian et al. (2019) completed a study is to determine clinical outcomes of patients that would have been scored falsely positive or falsely negative using the qSOFA scoring system. This is a retrospective cohort study that enrolled all adult patients over the age of 18 that were hospitalized with sepsis. The sample size was 21,191. Univariate and multivariate logistic regression analyses were used to identify independent risk factors associated with mortality. AUC was used to calculate the prognostic value of qSOFA. Sensitivity and specificity were calculated for false-positives and false-negatives. Finally, all categorical variables 40 were measured using chi-square. Limitations identified were using a database that was not originally designed for the study of qSOFA and qSOFA scores were calculated on the worst vital signs throughout admission and not upon presentation. The calculated sensitivity rate for the qSOFA tool was 50.2% and specificity rate was 78.1%. This supports other studies that the qSOFA has a high specificity rate. The authors suggest further prospective studies be completed to validate their findings. This subtheme highlights that when the qSOFA score is positive, there is a high potential that patient is septic. At a minimum, the patient needs urgent evaluation for another emergency. This evidence supports the initiation of the qSOFA tool in this particular E.D. setting as they do not currently use any type of screening tool. Subtheme 3, Ease of Use Regarding qSOFA The third subtheme is the ease of use with qSOFA. It has been reviewed that many tools exist to screen for sepsis. Some of these tools are validated and others are facility specific, but none of them are as easy to use as the qSOFA making compliance easier to attain. The other tools require lab studies or have several questions that need answered and when a nurse is in triage seeing five to ten patients per hour, they do not have time for complex screening tools. For this reason, qSOFA was selected as the intervention for the project. Two studies will be reviewed in this subtheme. Kwak et al. (2018) completed a study to test the combination of the ESI score and qSOFA score to predict mortality. The ESI score is calculated at the practice site on every patient before a provider can see a patient to pick up for 41 their caseload. The same items needed to calculate qSOFA are also incorporated into the ESI score making the use of qSOFA easier in a triage setting. This is a retrospective study that used patient records for patients age 15 and older. The sample size was 43,748. Patients were then divided into four groups for ESI 1, ESI 2, ESI 3 & ESI 4/5. Statistical tests used to measure continuous variables were t-test and one-way ANOVA. Categorical variables were tested using the chisquare test. Limitations identified included the specific diagnosis of patients was not available, which means some qSOFA positive patients were quickly seen based on the positive score, but did not have sepsis, but rather had other critical etiology occurring. The results showed that when the qSOFA score was positive it correlated with a higher mortality for ESI scores 1, 2 & 3. Positive qSOFA plus ESI 1 improved mortality prediction from 14.7% (ESI only) to 20.4% (qSOFA plus ESI). Positive qSOFA plus ESI 2 improved mortality prediction from 2.7% (ESI only) to 11.3% (qSOFA plus ESI). Positive qSOFA plus ESI 3 improved mortality prediction from 0.4% (ESI only) to 2.3% (qSOFA plus ESI). No changes were noted with ESI levels 4/5 and qSOFA. These results not only support the use of this tool for this project from an ease-of-use perspective, but they could also enhance overall care in the E.D. setting as an incidental finding. The authors have suggested further studies should be completed related specifically to sepsis and ESI plus qSOFA. Singer et al. (2017) completed a study is to determine the qSOFA scoring tool ability to predict inpatient mortality. This is a retrospective study that used adult patients over the age of 18 that presented to the E.D. with sepsis. The sample size was 22,530. The chi-square test was used to measure all categorical 42 variables. t-test and ANOVA were used to compare continuous variables. AUC was used to measure sensitivity and specificity. Limitations identified included the study was using inpatients who are sicker, thus qSOFA may have performed higher than expected since the acuity of the patient was higher. The results showed a positive qSOFA score of 2 was associated with a 12.8% mortality and qSOFA score of 3 was associated with 25% mortality. Sensitivity for qSOFA of 2 was 29% and specificity for qSOFA of 2 was 97%. These results validate prior studies of sensitivity and specificity, but the authors do make a point to explain how easy the scoring tool is to complete leading to increased compliance with usage. Future questions for research include further testing of qSOFA on E.D. only patients as the original study did not use any E.D. patients. This subtheme reveals how using the qSOFA tool is not difficult. While the tool may not be as sensitive as many would like, it does have a high specificity rating. Since the tool only needs three things to determine a score and these items are already assessed in triage to determine the ESI score, this tool aligns well with the project. Summary of Literature Review To conclude, a comprehensive literature review needs to be completed prior to the implementation of any project. This allows for the doctorate of nursing practice (DNP) prepared nurse to have a full understanding of the evidence that has been published and gives insight into how interventions will work within one’s respective practice site. This literature review supports the use of the qSOFA tool at the author’s practice site, as it has been identified that the organization elected to not use a screening methodology after the recommendation was made to stop using the SIRS tool. As described, this tool does have 43 a low to moderate sensitivity, but a high specificity. Given that the practice site is not using any screening process, even a low sensitivity tool is expected to produce an improvement. Additionally, the tool is easy to use, which makes compliance easier. Other tools such as NEWS which have better sensitivity and specificity are also more difficult to use for a site that is new to sepsis screening, which was documented in the literature that compliance with using the tool was not always 100%. No tool will work if it is not used. Understanding all these concepts leads to supporting the use of qSOFA for the practice improvement project. 44 Chapter 3: Methodology To review, the focus of this project is to improve the identification of potential sepsis patients in the emergency department setting. This will be accomplished by implementing the qSOFA sepsis screening tool during the triage phase of care in the emergency department. The facility has a goal of 30 minutes to collect a lactic acid, but they are currently not meeting that goal. This will serve as the metric that is measured to demonstrate how the qSOFA screening tool improved this collection time. This chapter will further explain the data collection procedures that will be used to measure statistical significance of the project. Statement of the Problem Current practice at the facility is not using any type of sepsis screening allowing sepsis to be found later during care and delaying treatment. The project will improve the collection time of the lactic acid level, which is one of the Sepsis-1 metrics. Reducing the time of lactic acid collection will allow for faster recognition of a sepsis condition and subsequent treatment. Clinical Question The following clinical question guides this quantitative project: Q1: To what degree does the implementation of The University of Pittsburgh quick sequential organ failure sepsis screening impact Centers for Medicare & Medicaid Services Sepsis-1 door to lactic acid collection when compared to current practice among adult patients arriving through triage or by ambulance in an urban emergency department in Pennsylvania over four-weeks? 45 Variable 1: qSOFA screening tool is the independent variable that is a scoring system identifying patients as either positive or negative making the scoring a nominal variable. Variable 2: Door to lactic acid measurement is a time measurement starting at zero minutes and working its way up. Having an absolute zero point makes this a ratio variable. Variable Variable Type qSOFA screening tool Independent Level of Measurement Nominal Door to lactic acid measurement Dependent Ratio The variables will be a within-subjects (or repeated-measures) design as the same data is in all collections both before and after the intervention. This design was selected, so that an accurate comparison to pre-implementation and post-implementation can be reviewed for statistical significance. This gives an objective view to how well the tool performed to improve the metric. Project Methodology The project used a quantitative methodology. Melnyk and Fineout-Overholt (2019) define a quantitative methodology as a study that examines the relationships between two more variables. Quantitative methodology can have several types of designs such as experimental design, quasi-experimental design, causal-comparative design, or time-series design, which will be discussed in the next section (Houser, 2018). 46 Quantitative methodology is appropriate when answering questions such as what is the problem, frequency of the problem or will this intervention improve the problem (Melnyk & Fineout-Overholt, 2019)? This project answers these questions in that the problem is poor compliance with sepsis metrics and a literature review that recommends using a screening intervention to improve compliance. This project does not measure how staff feel about using a sepsis screening tool or how patients react to being assessed for the clinical data being obtained, which would then define this as a qualitative methodology. Project Design The research question for this project calls for the introduction of an intervention to be applied to all patients. This means there is not a control group or a group that does not receive an intervention. Melnyk and Fineout-Overholt (2019) explain that the quasiexperimental design is one that does not have a control group, because random assignment is often not possible due to ethical concerns. In the case of this project, withholding the application of sepsis screening to random patients may cause harm by delaying care. The sample that was investigated was all adult patients presenting to the emergency department. The process to collect the data was the nurse labeling a paper qSOFA form that is positive and placing it in a bin to be collected, counted, and scanned into the EHR. This needed be a paper process as this tool is not built into the EHR. Population and Sample Selection The setting for the population was an urban E.D. that sees an annual volume of around 55,000. The population used for the study was the adult population age 18 and over that arrived by either private vehicle or ambulance. The qSOFA screening tool was used during the triage phase of care, which is the first phase of care. During this phase of 47 care, the nurse assessed for the criteria included in the qSOFA tool. If the patient met criteria, a paper qSOFA form was be completed and scanned into the EHR. No consent was needed for patients as the same clinical information being collected during the triage phase to assign an ESI score is the same information needed to collect for qSOFA. Patients were not be subjected to any different processes nor did they need to endure any undue hardship as a result of the quality improvement project. All patient confidentiality measures that are currently in place in the facility were maintained throughout the quality improvement project. Instrumentation or Sources of Data This project used data collected from the EHR and from a paper qSOFA tool. The use of a paper tool was be required due to cost and ease of collection. The data collected included time, gender, patient medical record number (MRN), zip code, qSOFA score and door to lactic acid collection time. A health insurance portability and accountability act (HIPAA) waiver was not be sought as patients were not contacted for permission to use their data and patient identifier data was not be displayed in the results. qSOFA Tool Details of the qSOFA tool will be discussed here. The qSOFA tool is a threequestion tool (Appendix B). The way to score points with qSOFA is to record one point for each abnormal area. The three areas are mental status, systolic blood pressure and respiration rate. To score a point for mental status the patient must have either a GCS of 14 or less or have a more pronounced altered state from their baseline for patients with underlying cognitive conditions such as dementia. To score a point for blood pressure, the systolic pressure must be less than 100. To score a point for respiration rate, the rate must be greater than 22. If a patient arrived with a GCS of 15, blood pressure of 118/62 48 and respiration rate of 18, their qSOFA score would be zero. If a patient arrived with a GCS of 14, blood pressure of 92/50 and respiration rate of 20, their score would be two. Scores of two or more are considered positive. Validity Sylvia and Terhaar (2018) define validity as an instrument that measures what it is intended to measure. To ensure validity for this project, this project followed the intended guideline and use for the qSOFA tool, which was to identify potential sepsis patients outside of the ICU setting. The tool was not being altered and any way and was only used on patients in the E.D. setting. By doing this, it preserves the integrity and validity of the tool as it was intended. In order to support internal validity, the project was been designed to implement one intervention (independent variable) and measure one metric (dependent variable). This was to reduce the possibility of having other variables that could impact the outcome. The project had an expectation to add to the external validity of the qSOFA tool by adding the generalizability to the out of ICU population. Reliability Sylvia and Terhaar (2018) define reliability as an instrument that yields the same results through repeated administrations of the test. Reliability has been established throughout the literature with an understanding of low to moderate sensitivity and high specificity. This information was reviewed in chapter two under the literature review. Therefore, this project implemented the qSOFA tool using the same out of ICU population with an expectation of the same results. 49 Data Collection Procedures To collect data, the project used data already being collected by the practice site for sepsis metrics by utilizing the certified and trained sepsis abstractors that are employed to abstract sepsis data for the practice site. The abstractors are required to have completed training approved by CMS to accurately abstract sepsis data using a standardized abstraction tool (CMS, 2021). Additionally, a quarterly sample of their abstractions are sent to a third-party vendor to be verified for accuracy. As this data collection process is already in place, the project continued to use this resource. The additional step of collecting the paper qSOFA forms was the responsibility of the emergency department quality assurance nurse, so the data could be counted and quantified to be used specifically for the project. This occurred daily. The population targeted was all adult patients over the age of 18 entering the emergency department. Only positive patients had a qSOFA form completed. For this data collection to begin for the project the organization and university required institutional review board (IRB) approval (Appendix C & D). The following items had to be completed to obtain approval. • CITI training completion • Practice site informed consent training • The qSOFA tool being used uploaded for review Data Analysis Procedures The descriptive statistics used for this project include mean, median and standard deviation for all variables. The independent variable was measured as yes or no. The dependent variable data was displayed in minutes. Inferential statistics needed to test the independent variable was be the t-test to test the differences in two related groups (pre- 50 intervention vs. post-intervention). The results were used to determine if any statistical change was noted to draw inference from. The data for this project was collected and the raw numbers were converted into mean, median and standard deviation by using frequencies to make the numbers meaningful. The data included registration time and lab collection time. To do this, first the data was abstracted and categorized accordingly. Next descriptive statistics were used to obtain the mean, median and standard deviation. Using a parametric test called the ttest, the difference between the pre-intervention group and post-intervention group was able to be calculated with the independent variable to determine statistical significance. A p-value of <0.05 was used to determine statistical significance. Potential Bias and Mitigation Potential bias that existed for this project was if a nurse decided to not screen a patient or a sample size that is not adequate. To mitigate these problems, education was completed with nearly every nurse, so they had an understanding on how to use the tool and why it is important to screen for sepsis. This did not guarantee elimination of bias, but did mitigate it. Sampling bias was mitigated by not randomly selecting patients and applying the screening tool to all adult patients entering the E.D. Finally, the data was evaluated by a statistician to eliminate a data analysis bias. Ethical Considerations Ethical considerations for the project evolved around data integrity. Since the same clinical data was collected is the same data currently collected in triage, no additional strain occurred on a patient. Additionally, this intervention did not involve any type physical task, it was merely an assessment. However, data including information about the patient such as an MRN, date and time of service and gender 51 was collected giving rise to the possibility that someone could be identified. Understanding this, management of the data followed the practice site data management plan and was also reviewed by the practice site IRB. Since the data being used was already captured by employees at the practice site, the needed safeguards were already in existence in the form of secured and encrypted servers and data warehouses. Additionally, any data being shared with the school in the form of the project was only minimally necessary data only such as time metric results. Vital signs information and MRN information was not shared in the results. Limitations Broad limitations were discussed in chapter one, but a more detailed review of limitations will be presented here. The qSOFA tool has its own set of limitations documented in the literature. It is known that some sepsis patients will be missed and the recognition of sepsis will have to occur via physical exam and clinical prudence. The data collection process was a manual process, which opened the opportunity for human error by not completing the screening or misunderstanding of the tool resulting in either a false negative or false positive score. Since this tool was only used on all adult patients, it was possible for intoxicated patients with a high respiration rate to be screened positive. Further clinical assessment and decision making will need to occur to determine if they are truly a septic patient. While these limitations were present, mitigation initiatives such as education, the use of established sepsis order sets and daily check in to assess progress and answer questions was completed. To better help with the change, the use of a change model was also used. Kotter (1996) developed his eight-step change model to assist in making change that is sustainable. These steps include developing a sense of urgency, establishment of a team, 52 developing vision, communicating the vision, empowering the staff to make the change, celebrating short-term wins, consolidating the wins to show improvement, and finally anchoring the change as daily culture and practice. These steps had already begun and will continued to be touched on not only before the implementation, but also during the implementation. Summary To summarize chapter three, this project answered the clinical question by implementing an evidence-based tool into practice and measuring the outcome and results. The project was a quantitative methodology with a quasi-experimental design, so that the intervention was applied to all adult patients entering the E.D. Data collection occurred using a paper qSOFA tool that was completed with a positive screen and the initiation of a sepsis order set. Descriptive statistics were used to determine statistical significance using the door to lactic acid collection time as the metric determining statistical significance. Next will be chapter four and a review of the data analysis and results. 53 Chapter 4: Data Analysis and Results This chapter will review the data analysis and results from the project. This review will also include the sample comparisons both pre- and post-intervention as well as explanations about the data. Finally, this chapter will review the statistical analysis used to answer the PICOT question. All statistical analysis was completed using Microsoft® Excel and Jamovi Statistics, version 2.3 (2022). Descriptive Data The descriptive data used to compare the samples were gender and age. The preimplementation sample was representative from January 2022 to July 2022. It consisted of 197 patients with an ICD-10 code of sepsis, however eight cases were missing collection times and needed to be removed bringing the sample size to 189. The postimplementation data consisted of all qSOFA positive patients with a sample size of 50. Pre- and post-implementation gender related data were compared to ensure comparable samples. Chi square was calculated and noted no significant difference between samples (see Table 1). 54 Table 1 Comparison of Gender Pre/Post qSOFA gender PRE/POST F M Total PRE 92 97 189 POST 24 26 50 Total 116 123 239 X2 (1, N=239) = 0.01, p = .932 Pre and post implementation age related data was compared to ensure comparable samples. t-test was used to determine no significant difference in age related data was found (see Table 2). 55 Table 2 Comparison of Ages Pre/Post qSOFA Age Group N Mean Median SD PRE 189 69.1 70.0 14.6 POST 50 66.6 69.5 17.5 t(237) = 1.0, p = .316 Comparison of Minutes to Lactate Pre and Post qSOFA Implementation Review of the clinical question is important to better understand this section. To what degree does the implementation of The University of Pittsburgh quick sequential organ failure sepsis screening impact Centers for Medicare & Medicaid Services Sepsis-1 door to lactic acid draw collection when compared to current practice among adult patients arriving through triage or ambulance in an urban emergency department in Pennsylvania over four-weeks? As stated before, the original sample was an n=197 and eight of these cases were missing a collection time when completing a retrospective record review. These cases needed to be removed. The post-implementation sample was from October 2022 and consisted of 50 patients that were screened with a positive qSOFA screening. The eight cases in the retrospective data that were missing collection times were removed leaving an n=189 pre-intervention and n=50 postintervention. The comparison of minutes from registration time to lactic acid collection for both pre- and post-implementation was compared next. Pre-implementation data included 189 patients that had an ICD-10 code of sepsis. No sepsis screening was being completed 56 during this time. Post-implementation data included 50 patients that were all qSOFA positive, but final end-coding with ICD-10 codes has not been complete at the time of writing this manuscript. The statistical test used in the analysis was the t-test. Assumptions for the t-test are independence, normality and equal variances. Independence meaning both groups were independent of each other was evaluated and both groups were found to be independent. The t-test assumption for independence was met. Normality assumption was evaluated using the Shapiro-Wilk normality test (See Table 3). Table 3 Normality Test (Shapiro-Wilk) – All Cases MINUTES W p 0.601 < .001 Note. A low p-value suggests a violation of the assumption of normality Based on Shapiro-Wilk test the data does not meet the normality assumption. In order to improve normality of the data, eight data points in the pre-data were excluded as outliers. These outliers had lactic acid collection times greater than 350 minutes and of note, all were qSOFA negative. As a result, the pre-implementation sample size will be 181 and the post-implementation sample size will be 50. The Shapiro-Wilk test was repeated with outliers removed and was found to have not met normality, but was noted to be improved. Fien et al. (2022) explains the dependent variable should be normally or near-normally distributed for each group. The t-test is robust for minor violations in normality (Fien et al., 2022). Given this information, the decision was made to use the ttest for the analysis (See Table 4). 57 Table 4 Normality Test (Shapiro-Wilk) MINUTES W p 0.839 < .001 Note. A low p-value suggests a violation of the assumption of normality The third assumption for t-test is equal variances. The homogeneity of variance test was completed (Levene’s Test). See table 5. Table 5 Homogeneity of Variances Test (Levene's) MINUTES F df df2 p 10.0 1 229 0.002 Note. A low p-value suggests a violation of the assumption of equal variances Homogeneity of variance assumption was not met. However, to deal with the violation of the homogeneity the Welch’s t-test will be used (Glen, 2022). Results A t-test with outliers removed was completed to answer the PICOT question. The null hypothesis is the average minutes pre-intervention equals average minutes postintervention. Table 7 shows the results of the t-test. The results of the t-test demonstrated a statistically significant difference (t(137) = 2.43, p = 0.017). This rejects the null hypothesis and accepts the alternative hypothesis that there is a difference between preand post-minutes. Finally, the Cohen’s d was used to determine the effect size. Fien et al. (2022) defines a small, medium, and large effects for the t-test to be .2, .5 and .8 respectively. The Cohen’s d result was 0.329 revealing a small to medium effect (see 58 Table 6). Table 6 Independent Samples T-Test Statistic df p 137 0.017 Effect Size MINUTES Welch's t 2.43 Cohen's d 0.329 ᵃ Levene's test is significant (p < .05), suggesting a violation of the assumption of equal variances Summary To summarize the data, the t-test was the parametric test used to determine statistical significance. However, in order to better normalize the distribution of the data, eight outlier cases needed to be removed from the pre-intervention data. Doing this allowed for a more normal distribution of the data. The results using Welch’s t-test and was found to have a p-value of 0.017 demonstrating statistical significance between the pre-intervention data and post-intervention data with a Cohen’s d indicating a small to medium effect. 59 Chapter 5: Summary, Conclusions, and Recommendations After completion of the project and analysis of the data, this chapter will review the summary of the project, conclusion of results and recommendations for future action. As previously reviewed in chapter four, the project did yield statistical significance, but more importantly the project exhibited clinical significance and improvement in the clinical management of the septic population in this E.D. Summary of the Project The sepsis screening project was focused on the implementation of the qSOFA sepsis screening tool and determination if that single action would result in improvement of the lactic acid collection time. Several tools were evaluated to be used and some were also noted to have better outcomes, but higher complexity as it related to implementation. For this reason, the qSOFA tool was chosen when considering the post-pandemic staffing situation, increased observation of boarding patients in the E.D. and leadership turnover within the department. Implementation of a more robust, but complex tool would risk non-compliance. Critical to the project was staff acceptance. The nursing staff felt the tool was very simple to use and recognized its value when positive that something was clinically wrong with the patient and the patient needed to be brought to a provider quickly. Many of the nursing staff likened a positive screen to a positive stroke or STEMI situation to at least get the appropriate labs ordered and properly leveled using the ESI system. The provider staff also found value in having positively screened patients either presented directly to them or flagged with a high ESI level for expedient pick up. Additionally, mid-level providers in triage found use in taking the qSOFA positive patient and being able to call the charge nurse or attending physician to quickly room the 60 patient and begin their clinical workups with sepsis being a differential diagnosis. In many cases, this resulted in bypassing a three-hour wait in the waiting room. Summary of Findings and Conclusion The answer to the PICOT question, the implementation of the qSOFA tool did yield statistical significance in the collection times of the lactic acid. As described in the literature, the qSOFA tool did demonstrate a low sensitivity and high specificity for this project as well. Moreover, the project did expand upon the scientific knowledge of the staff related to sepsis knowledge and management. As discussed in the themes, knowledge deficit regarding sepsis is a significant concern. By virtue of the project, nearly the entire nursing staff were able to be educated on the basics of sepsis pathophysiology and understand the treatments for sepsis. Additionally, the project answered many of the “whys” that staff may have had regarding the three-hour bundle and six-hour bundles. The project also recognized clinical significance with 23 patients that screened positive and were found to have other clinical etiology that needed to be addressed urgently. The ability to take purposeful action based on the positive screen brought a sense of satisfaction to many of the staff as well. This is largely in part, because it was not just their “gut reaction” that something was not right, but they had an objective positive score to back their findings and move the patient to the next phase in care. This makes conversations between attending physicians and nurses easier too. Finally, this project also brought forth the value in accurate measurement of respiratory rates. It was found during the retrospective review of the sepsis patients that many had documented respiratory rates of 16 or 18, but consistently monitored respiratory rates of 25 from the bedside monitor. The project was able to address this clinical measurement concern and 61 provide an opportunity to educate nursing staff regarding the significance of elevated respiratory rate in the presence of acidosis. While the project demonstrated statistical significance related to the PICOT question, it also exposed many of the underpinning short falls needed to have a successful sepsis program. This ended up being the most significant part of the project and demonstrates the importance of understanding the culture and education levels of the staff that will be tasked with identification and caring of these patients. Now at the completion of this project, the path forward is much clearer. Implications The implications this project had surrounded the importance of using evidencebased practice to guide clinical actions. Prior to initiation of the screening tool, hemodynamically stable sepsis patients were simply triaged, evaluated according to ESI level and time of arrival. However, now that an evidence-based process has been implemented (sepsis screening), these patients are being called out and moved to the front of the line for timely evaluation and treatment. Teamwork was another implication that evolved from this project. Emergency medicine is a team sport, but so is sepsis management. The project facilitated the focus on sepsis in the E.D., but it also spilled over to include the hospitalist and to some extent, the inpatient nursing staff. The final implication from this project was trust. The staff began to learn to strengthen their trust in each other when they had a positive patient and trust in the tool itself when they saw a positive patient achieve all the three-hour bundle metrics. This is clearly an important factor when working in a E.D. setting. 62 Theoretical Implications As reviewed in chapter two, the project was following King’s system theory as the lack of sepsis screening was identified as a systems issue. The use of this theory turned out to be quite valuable, because it forced cultural change for the whole organization. Not only did the project change the culture of the E.D., it also received attention from the executive leadership team and the quality department, which resulted in reestablishing the hospital-wide sepsis committee. Furthermore, this project allowed for sepsis management to be placed on performance dashboards at the physician and nursing levels, so sustainability and expectations of continued improvement can occur. Recommendations The first recommendation would be the implementation of a screening tool that was incorporated in the EHR. Automation and leveraging technology would hypothetically improve sepsis recognition and early management. While this practice site did note improvement in their sepsis data using the qSOFA tool, it would be wise to explore other screening tools based on recent literature (Evans et al., 2021). The most significant recommendation from this project is to establish a sepsis alert process. While the data point of order set entry was not measured, the quality department did note improvement in the launching of sepsis order sets early during care. Even though improvement in collection times was noted, more improvement could be discovered if an actionable process was put into place. This idea is already being discussed at the sepsis committee and is part of the sustainability of the project. The final recommendation would be the use of the qSOFA tool with the local paramedic providers. Identification of a potential sepsis patient in the pre-hospital environment has potential to be impactful. The qSOFA would be an easy tool for pre- 63 hospital providers to learn and use. After a sepsis alert process is established internally, the pre-hospital staff could be activating that process before the patient arrives thus improving the timeliness of interventions. Conclusion To conclude, low performance in CMS sepsis metrics and clinical outcomes was identified as a clinical problem that needed to be addressed. Through literature review, it was found that the use of a sepsis screening tool was one of the primary methods to resolve this problem. Since the practice site did not use any type of sepsis screening methodology, it became clear that this was a systems issue that created and supported the culture of no screening. King’s systems theory was used to build the framework of the project and it was decided to implement an easy-to-use sepsis screening tool to assist in changing the culture of the department. The qSOFA tool was the best fit for this situation to help change the culture, educate staff, and begin to make improvements in sepsis management. Post-implementation, both clinical and statistical significance was noted in the management of the septic population. Additionally, clear next steps emerged as a result of the project to reinforce sustainability and continuous improvement. This entire project demonstrates the importance of finding evidence-based practices that can be implemented to improve actual clinical outcomes. 64 References Abdullah, S.M., Grand, J., Sijapati, A., Puri, P.R. & Nielson, F.E. (2020). qSOFA is a useful prognostic factor for 30-day mortality in infected patients fulfilling the SIRS criteria for sepsis. American Journal of Emergency Medicine, 38(3), 512-516. https://doi.org/10.1016/j.ajem.2019.05.037 Aksu, A., Gulen, M., Avci, A. & Satar, S. (2018). Adding lactate to SOFA and qSOFA scores predicts in-hospital mortality better in older patients in critical care. European Geriatric Medicine, 10(3), 445-453. https://doi.org/10.1007/s41999-01900179-z Asai, N., Watanabe, H., Shiota, A., Kato, H., Sakanashi, D., Hagihara, M., Koizumi, Y., Yamagishi, Y., Suematsu, H. & Mikamo, H. (2019). Efficacy and accuracy of qSOFA and SOFA scores as prognostic tools for community-acquired and healthcare-associated pneumonia. International Journal of Infectious Diseases, 84(7), 89-96. https://doi.org/10.1016/j.ijid.2019.04.020 Barbara, P., Graziano, C., Caputo, W., Litvak, I., Battinelli, D. & Hahn, B. (2018). The quick sequential organ failure assessment (qSOFA) identifies septic patients in the out-of-hospital setting. American Journal of Emergency Medicine, 36(6), 10221026. https://doi.org/10.1016/j.ajem.2018.01.073 Baumann, B.M., Greenwood, J.C., Lewis, K., Nuckton, T.J., Darger, B., Shofer, F.S., Troeger, D., Jung, S.Y., Kilgannon, J.H. & Rodriguez, R.M. (2020). Combining qSOFA criteria with initial lactate levels: Improved screening of septic patients for 65 critical illness. American Journal of Emergency Medicine, 38(5), 883-889. https://doi.org/10.1016/j.ajem.2019.07.003 Brink, A., Alsma, J., Verdonschot, R.J., Marie Rood, P.P., Zietse, R., Lingsma, H.F. & Elisabeth Schuit, S.C. (2019). Predicting mortality in patients with suspected sepsis at the emergency department; A retrospective cohort study comparing qSOFA, SIRS and national early warning score. PLOS One, 14(1), https://doi.org/10.1371/journal.pone.0211133 Burnham, J.P. & Kollef, M.H. (2018). qSOFA score: Predictive validity in Enterobacteriaceae bloodstream infections. Journal of Critical Care, 43, 143-147. https://doi.org/10.1016/j.jcrc.2017.09.011 Canet, E., Taylor, D.M., Khor, R., Krishnan, V. & Bellomo, R. (2018). qSOFA as a predictor of mortality and prolonged ICU admission in emergency department patients with suspected infection. Journal of Critical Care, 48, 118-123. https://doi.org/10.1016/j.jcrc.2018.08.022 Centers for Medicare and Medicaid Services (2021). Abstraction resources. Retrieved from https://qualitynet.cms.gov/inpatient/data-management/abstraction-resources Costa, R.T., Nassar, A.P. & Caruso, P. (2018). Accuracy of SOFA, qSOFA and SIRS scores for mortality in cancer patients admitted to an intensive care unit with suspected infection. Journal of Critical Care, 45, 52-57. https://doi.org/10.1016/j.jcrc.2017.12.024 Delawder, J.M. & Hulton, L. (2020). An interdisciplinary code sepsis team to improve sepsis-bundle compliance: A quality improvement project. Journal of Emergency Nursing, 46(1), 91-98. https://doi.org/10.1016/j.jen.2019.07.001 66 Evans, L., Rhodes, A., Alhazzani, W., Antonelli, M., Coppersmith, C.M., French, C., Flavia, R., Mcintyre, L., Osterman, M., Prescott, H.C., Shorr, C., Simpson, S., Wiersinga, J.W., Alshamsi, F., Angus, D.C., Arabi, Y., Azevedo, L., Beale, R., Beilman, G.,…Levy, M. (2021). Surviving sepsis campaign: International guidelines for management of sepsis and septic shock 2021. Critical Care Medicine 49(11), https://doi.org/10.1097/CCM.0000000000005337 Fein, E.C., Gilmour, J., Machin, T. & Hendry, L. (2022). Statistics for research students: An open access resource with self-tests and illustrative examples. https://usq.pressbooks.pub/statisticsforresearchstudents/ Fernando, S.M., Barnaby, D.P, Herry, C.L., Gallagher, J.E., Shapiro, N.I. & Seely, A.J. (2018). Helpful only when elevated: Initial serum lactate in stable emergency department patients with sepsis is specific, but not sensitive for future deterioration. Journal of Emergency Medicine, 54(6), 766-773. https://doi.org/10.1016/j.jemermed.2018.01.040 Filbin, M.R., Thorsen, J.E., Lynch, J., Gillingham, T.D., Pasakarnis, C.L., Capp, R., Shapiro, N.I., Mooncai, T., Hou, P.C., Heldt, T. & Reisner (2018). Challenges and opportunities for emergency department sepsis screening at triage. Scientific Reports, 8, https://doi.org/10.1038/s41598-018-29427-1 Gattinoni, L., Vasques, F., Camporota, L., Meessen, J, Romitti, F., Pasticci, I., Duscio, E., Vassalli, F., Forni, L.G., Payen, D., Cressoni, M., Zanella, A., Latini, R., Quintel, M. & Marini, J.J. (2019). Understanding lactatemia in human sepsis, potential impact for early management. American Journal of Respiratory and 67 Critical Care Medicine, 200(5), 582-589. https://doi.org/10.1164/rccm.2018122342OC Glenn, S. (2022). Welch’s ANOVA: Definition and Assumptions. Retrieved from https://www.statisticshowto.com/welchs-anova/ Goulden, R., Hoyle, M.C., Monis, J., Railton, D., Riley, V., Martin, P., Martina, R. & Nsutebu, E. (2018). qSOFA, SIRS, and NEWS for predicting in hospital mortality and ICU admission in emergency admissions treated as sepsis. Emergency Medicine Journal, 35(6), 345-349. http://dx.doi.org/10.1136/emermed-2017207120 Houser, J. (2018). Nursing research: Reading, using, and creating evidence. (4th ed.) Jones & Bartlett Learning. Husabo, G., Nilsen, R.M., Flaatten, H., Solligard, E., Frich, J.C., Bondevik, G.T., Braut, G.S., Walshe, K., Harthug, S. & Hovlid, Einar (2020). Early diagnosis of sepsis in emergency departments, time to treatment, and association with mortality: An observational study. PLoS One, 15(1), https://doi.org/10.1371/journal.pone.0227652 Jung, Y.T., Jeon, J., Park, J.Y., Kim, M.J., Lee, S.H. & Lee, J.G. (2018). Addition of lactic acid levels improves the accuracy of quick sequential organ failure assessment in predicting mortality in surgical patients with complicated intraabdominal infections: A retrospective study. World Journal of Emergency Surgery, 13(14), https://doi.org/10.1186/s13017-018-0173-6 Kievlan, D.R., Zhang, L.A., Chang, C.H., Angus, D.C. & Seymour, C.W. (2018). Evaluation of repeated qSOFA measurements among patients with suspected 68 infection. Critical Care Medicine, 46(12), 1906-1913. https://https://dx.doi.org/10.1097%2FCCM.0000000000003360 King, I. (1968). A conceptual frame of reference for nursing. Nursing Research, 17(1), 27-30. Kotter, J.P. (1996) Leading Change. Harvard Business Press. Kwak, H., Suh, G.S., Kim, T., Kwon, W.Y., Kim, K.S., Jung, Y.S., Ko, J.K. & Shin, S.M. (2018). Prognostic performance of emergency severity index (ESI) combined with qSOFA score. American Journal of Emergency Medicine, 36(10), 1784-1788. https://doi.org/10.1016/j.ajem.2018.01.088 LeGuen, M. Ballueer, Y., McKay, R., Eastwood, G., Bellomo, R. & Jones, D. (2018) Frequency and significance of qSOFA criteria during adult rapid response team reviews: A prospective cohort study. Resuscitation, 122, 13-18. https://doi.org/10.1016/j.resuscitation.2017.11.036 Mecham, I.D., Dean, N.C., Wilson, E.L., Jephson, A.R. & Lanspa, M.J. (2018). The association of qSOFA, SOFA and SIRS with mortality in emergency department pneumonia patients. University of Louisville Journal of Respiratory Infections, 2(2), 12-18. https://10.18297/jri/vol2/iss2/4 Melnyk, B.M. & Fineout-Overholt, E. (2019). Evidenced-based practice in nursing and healthcare. (4th ed.) Wolters Kluwer. Meyer, N.M., Harhay, M.O., Small, D.S., Prescott, H.C., Bowles, K.H., Gaieski, D.F. & Mikkelsen, M.E. (2018). Temporal trends in incidence, sepsis-related mortality, and 69 hospital-based acute care after sepsis. Critical Care Medicine, 46(3), 354-360. https://https://dx.doi.org/10.1097%2FCCM.0000000000002872 Mitzkewich, M. (2019). Sepsis screening in triage to decrease door-to-antibiotic time. Journal of Emergency Nursing, 45(3), 254-256. https://doi.org/10.1016/j.jen.2018.08.002 Moore, W.R., Vermuelen, A., Taylor, R., Kihara, D. & Wahome, E. (2019). Improving 3hour sepsis bundled care outcome: Implementation of a nurse-driven sepsis protocol in the emergency department. Journal of Emergency Nursing, 45(6), 690-698. https://doi.org/10.1016/j.jen.2019.05.005 Ortega, R.N., Rosin, C., Bingisser, R. & Nickel, C.H. (2019). Clinical scores and formal triage screening of sepsis and adverse outcomes on arrival in an emergency department all-comer cohort. Journal of Emergency Medicine, 57(4), 453-460. https://doi.org/10.1016/j.jemermed.2019.06.036 Perman, S.M., Mikkelsen, M.E., Goyal, M., Ginde, A., Bhardwaj, A., Drumheller, B., Sante, S.C., Agarwal, A.K. & Gaieski, D.F. (2020). The sensitivity of qSOFA calculated at triage and during emergency department treatment to rapidly identify sepsis patients. Scientific Reports, 10, https://doi.org/10.1038/s41598-020-77438-8 Rodriguez, R.M., Greenwood, J.C., Nuckton, T.J., Darger, B., Shofer, F.S., Troeger, D., Jung, S.Y., Speich, K.G., Valencia, J., Kilgannon, J.H., Fernandez, D. & Baumann, B.M. (2018). Comparison of qSOFA with current emergency department tools for 70 screening of patients with sepsis for critical illness. Emergency Medicine Journal, 35(6), 350-356. http://dx.doi.org/10.1136/emermed-2017-207383 Rosenqvist, M., Bengtsson-Toni, M., Tham, J., Lanbeck, P., Melander, O. & Akesson, P. (2020). Improved outcomes after regional implementation of sepsis alert: A novel triage model. Critical Care Medicine, 48(4), 484-490. https://doi.org/10.1097/CCM.0000000000004179 Seymour, C., Liu, V., Iwashyna, T., Brunkhorst, F.M., Rea, T.D., Scherag, A., Rubenfeld, G., Kahn, J.M., Shankar-Hari, M., Singer, M., Deutschman, C.S., Escobar, G.J. & Angus, D.C. (2016). Assessment of clinical criteria for sepsis for the third international consensus definitions for sepsis and septic shock (sepsis-3). JAMA, 315(8), 762-774. https://doi:10.1001/jama.2016.0288 Shah, T., Sterk, E. & Rech, M.A. (2018). Emergency department sepsis screening tool decreases time to antibiotics in patients with sepsis. American Journal of Emergency Medicine, 36(10), 1745-1748. https://doi.org/10.1016/j.ajem.2018.01.060 Shetty, A.L., Thompson, K., Byth, K., Macaskill, P., Green, M., Fullick, M., Lander, H. & Iredell, J. (2018). Serum lactate cut-offs as a risk stratification tool for in-hospital adverse outcomes in emergency department patients screed for suspected sepsis. BMJ, 8(1), http://dx.doi.org/10.1136/bmjopen-2016-015492 Singer, A.J., Ng, J., Thode, H.C., Spiegel, R. & Weingart, S. (2017). Quick SOFA scores predict mortality in adult emergency department patients with and without 71 suspected infection. Annals of Emergency Medicine, 69(4), 475-479. https://doi.org/10.1016/j.annemergmed.2016.10.007 Sylvia, M.L. & Terhaar, M.R. (2018). Clinical Analytics and Data Management for the DNP. (2nd ed.) Springer Publishing. The Jamovi project (2022). Jamovi (Version 2.3) [Computer Software]. Retrieved from https://www.jamovi.org Tian, H., Zhou, J., Weng, L., Hu, X., Peng, J., Wang, C., Jiang, W., Du, X., Xi, X., An, Y., Duan, M. & Du, B. (2019). Accuracy of qSOFA for the diagnosis of sepsis-3: A secondary analysis of a population-based cohort study. Journal of Thoracic Disease, 11(5), 2034-2042. https://dx.doi.org/10.21037%2Fjtd.2019.04.90 Tokioka, F., Okamoto, H., Yamazaki, A., Itou, A. & Ishida, T. (2018). The prognostic performance of qSOFA for community-acquired pneumonia. Journal of Intensive Care, 6(46), https://doi.org/10.1186/s40560-018-0307-7 72 Appendix A 73 Appendix B 74 Appendix C 75 Appendix D