ASSIGNMENT 1

EFFICACY OF THE IMPLEMENTATION OF EARLY SEVERE SEPSIS STRATEGIES ON A MEDICAL SURGICAL UNIT

Presented in Fulfillment of the Requirements for the Degree of

Doctor of Nursing Practice

Nova Southeastern University Health Professions Division

College of Nursing

Jorge Hirigoyen 2015

Copyright by Jorge Hirigoyen, 2015

All Rights Reserved

v

Abstract

Background: Septicemia is among the leading causes of death in the U.S. Severe sepsis has been largely studied in the intensive care unit. Limited research is available regarding sepsis identification and treatment in medical-surgical units. Purpose: The purpose of this capstone study was to evaluate the efficacy of a severe sepsis bundle in a medical surgical non-ICU, specifically focusing on reducing sepsis transfer to higher levels of care and sepsis hospital length of stay. The sensitivity and specificity of the severe sepsis tool was also evaluated. Theoretical Framework: Donabedian’s model was used as a guiding framework for this project. Methods: A quasi-experimental pre-post comparison design was utilized to evaluate the sepsis transfers to higher level of care and sepsis hospital length of stay before and after the implementation of a severe sepsis bundle. Same approach was used to calculate the sensitivity and specificity of the tool. Results: The tests revealed that there was no statistical significance (p = .291) in terms of length of stay between the pre- and post-implementation groups. The study revealed that in fact there was a decrease in sepsis length of stay, from pre implementation of 8.53 days to post implementation of 7.26 days. Evaluation of sepsis transfers to higher level of care revealed that during the 6-month period of the study, there were zero transfers of septic patients. In regards to the sensitivity and specificity of the tool, the study yielded a sensitivity of 55.85% and a specificity of 99.54%. The positive predictive value (PPV) of the tool was estimated at 95.83%, negative predictive value (NPV) was estimated at 92.83% and disease prevalence was 15.95%. Area under the receiver operating curve was 0.777. Conclusions: The study’s results could have been obtained due to multiple limitations to the study. These limitations could be potential opportunities for future research. Nevertheless, the study demonstrated that early identification and early interventions of sepsis bundles are effective in delivery of evidence-based care. The severe sepsis bundle should be incorporated into all medical surgical wards within the organization and even be adapted to other sister institutions.

vi

Acknowledgments

I would like to thank the following people:

I would like to thank my Capstone committee: Dr. C. Christine Orton, my chairperson,

Dr. Diane Whitehead, and Dr. Andra Hanlon for their continued guidance and support

that with much determination were able to guide me in the right direction with their

insightful comments and words of encouragement.

I would like to express my sincere gratitude to my mentor Jill M. Szymanski for the

continuous support of my DNP study and related research, for her patience, motivation,

and immense knowledge. Her guidance led me throughout the development,

implementation, and writing of this capstone. I could not have imagined having a better

advisor and mentor for my DNP study.

I would like to also thank God among all things, my family: my parents and my sister,

colleagues, and staff at Baptist Hospital of Miami for their endless support, patience, and

eagerness, which provided me with the motivation and desire to complete this project.

I would like to express special thanks to my wife Stephanie, for her never-ending support

during the completion of this project. I would have not been able to complete this

without her support, especially since she acted as my catalyst in beginning this journey.

Lastly, I would like to dedicate this degree to my two lovable children, Ryan and Elise;

they are everything to me.

Table of Contents Title Page .........................................................................................................................i Signature Pages .............................................................................................................. ii Copyright ....................................................................................................................... iv Abstract ........................................................................................................................... v Acknowledgements ........................................................................................................ vi Table of Contents ..........................................................................................................vii List of Tables.................................................................................................................. ix List of Figures ................................................................................................................. x

Chapter 1: Nature of Project and Problem Identification ..................................................1

Significance of the Problem .........................................................................................4 Problem Statement and Purpose ...................................................................................5 Project Objectives ........................................................................................................7 Theoretical Framework: Donabedian's Model ..............................................................8 Theoretical Framework as It Relates to Clinical Problem ........................................... 12 Significance of the Study ........................................................................................... 14

Healthcare Practice ................................................................................................. 14 Healthcare Outcomes .............................................................................................. 15 Healthcare Delivery ................................................................................................ 15 Healthcare Policy ................................................................................................... 17

Summary ................................................................................................................... 18

Chapter 2: Literature Review ......................................................................................... 19 Epidemiology ............................................................................................................. 20 Treatment Recommendations and Modalities ............................................................. 22 Implementation Strategies .......................................................................................... 25 Biomarkers for Sepsis Identification........................................................................... 27 Procalcitonin Biomarker as a De-Escalation of Antibiotic Therapy ............................ 32 Summary ................................................................................................................... 33

Chapter 3: Methods ....................................................................................................... 34 Project Design ............................................................................................................ 34 Setting ....................................................................................................................... 35 Inclusion/Exclusion Criteria ....................................................................................... 36 Ethical Considerations and Confidentiality ................................................................. 37

Waiver of Informed Consent................................................................................... 37 Risk and Benefits.................................................................................................... 38

Goals and Objectives ................................................................................................. 38 Project Phases ............................................................................................................ 39 Budget ....................................................................................................................... 43

Budget Analysis ..................................................................................................... 43 Information Technology ............................................................................................. 44

Research Department ................................................................................................. 44 Timeline of Project .................................................................................................... 44 Determination of Sample Size: Power Analysis .......................................................... 45 Statistical Tests .......................................................................................................... 46 Outcome Measures ..................................................................................................... 46 Validity and Reliability of the Sepsis Screening Tool ................................................. 48 Summary ................................................................................................................... 49

Chapter 4: Results and Discussion ................................................................................. 50 Descriptive Results .................................................................................................... 51

Objective 1 ............................................................................................................. 51 Objectives 2 and 3 .................................................................................................. 51 Objective 4 ............................................................................................................. 52 Statistical Results ................................................................................................... 53

Strengths and Limitations of the Project ..................................................................... 57 Strengths ................................................................................................................ 57 Limitations ............................................................................................................. 59

Implications of the Study ........................................................................................... 60 Implications for Nursing Practice ........................................................................... 60 Implications for Healthcare Delivery ...................................................................... 61 Implications for Healthcare Policy .......................................................................... 62 Future Research ...................................................................................................... 63

Summary ................................................................................................................... 64

References ..................................................................................................................... 66 Appendix A: Nova IRB Exempt Letter .......................................................................... 71 Appendix B: Letter of Project Support from Medical Surgical Unit ............................... 73 Appendix C: Paper Copy of Severe Sepsis Tool............................................................. 75 Appendix D: Proposed Severe Sepsis Algorithm ........................................................... 76 Appendix E: Sepsis Order Set in Medical Surgical Floor ............................................... 77 Appendix E: Sepsis Order Set in Medical Surgical Floor ............................................... 78 Appendix F: Education Flyer ......................................................................................... 79 Appendix G: International Classification of Diseases (ICD-9) Coding ........................... 80

List of Tables

Table 1 Description of Comparison Groups ................................................................. 52 Table 2 Independent t-Test Results for Each Group ..................................................... 54 Table 3 2x2 Tabulation Design .................................................................................... 56 Table 4 Area Under the Curve ..................................................................................... 56 Table 5 Sensitivity and Specificity ............................................................................... 57

List of Figures

Figure 1. Adaptation of structure, process, and outcome model (Donabedian, 1980). ..... 81 Figure 2. Receiver operating curve (ROC). ................................................................... 82

Chapter 1: Nature of Project and Problem Identification

Sepsis is the tenth leading cause of death in the United States. The number and

rate per 10,000 hospitalizations for sepsis or severe sepsis has more than doubled from

2000 to 2008, with incidence increasing approximately 5%-10% each year (Hall,

Williams, DeFrances, & Golosinskiy, 2011). Sepsis is also the leading cause of death in

the critical care population regardless of multiple innovations in diagnosing and

resuscitation therapy (Sankar & Webster, 2013). According to the Centers for Disease

Control and Prevention (CDC), hospitalization for severe sepsis as the primary diagnosis

increased from 326,000 in 2000 to 727,000 in 2008 and as a secondary diagnosis from

621,000 in 2000 to 1,141,000 in 2008 (Hall et al., 2011). Severe sepsis and septic shock

are not only a problem in the United States but also worldwide. This pandemic affects

millions of individuals every year worldwide, approximately one out of four individuals

diagnosed with sepsis or septic shock dies (Dellinger et al., 2013). Anderson and

Schmidt (2010) concluded, that worldwide there are approximately 18 million new cases

of sepsis each year, with a mortality rate range estimated about 30% to 60%. Recent

approaches have targeted both diagnosing sepsis and rapidly treating those septic

patients, thus preventing further decompensation. Accurately identifying sepsis at an

early stage can potentially decrease mortality by 5%-10%, consequently accentuating the

importance of early detection and prompting therapeutic treatment (Singer, 2013).

Connecting early sepsis recognition with pathogen identification allows for practitioners

to effectively administer proper antibiotic therapy and treatment more rapidly, thereby

2

decreasing mortality and morbidity (Singer, 2013). Sepsis can be defined according to

the Surviving Sepsis Campaign of 2012 as a systemic, poisonous host response to an

infected pathogen leading to severe sepsis (Dellinger et al., 2013). In order for an

individual to be considered septic, serial ordinal symptoms must be present including

systemic inflammatory response syndrome (SIRS). SIRS describes a systemic

inflammation resulting from any major insult to the body, leading to a specific criteria

and affecting either temperature, respiratory rate, white blood cell count, and/or heart

rate. Sepsis is identified by the presence of SIRS criteria along with a known or

suspected infection. Severe sepsis with organ failure occurs in the presence of sepsis

with one of the signs and symptoms of organ failure leading to septic shock, which is

defined as severe sepsis with hypotension without response to fluids resuscitation

(Dillinger et al., 2013).

End organ failure is extremely important in the topic of sepsis; the goal of

healthcare providers is to identify and treat sepsis before any organ dysfunction occurs.

Tromp et al. (2011) reported on their wide-ranging study on septicemia that organ failure

occurred in 19.1% of sepsis patients from 1979 to 1989 and 30.2% from 1990 to 2000. In

the same study, the number of patients who experienced organ failure more than doubled

from 2.7% to 7.1%, demonstrating that sepsis is on a steady incline (2011). Rohde et al.

(2013) conducted a study evaluating the infection rate and organ dysfunction in patients

with severe sepsis in non-ICU, the study concluded that organ dysfunction was associated

with an increased mortality rate by 41%, hypotension being the most common

dysfunction.

3

In a combined approach to improve the increasing rate and complications of

sepsis and severe sepsis, healthcare providers have formulated different methods to not

only rapidly identify sepsis but also treat it. As a result, the Surviving Sepsis Campaign

(SSC) was launched in 2004 (Tromp et al., 2011), and a set of guidelines was created for

sepsis management. Since the initial creation of the SSC in 2004, two newer versions

have been developed: one in 2008 and more recently in 2012. Within the guidelines

recommended by the SSC, the resuscitation bundle and the management bundle are

among the most important components for the treatment of septic patients (Dillinger et

al., 2013). Since 2004, the sepsis bundles have been implemented in ICUs, emergency

departments and step-down units but have failed to be adopted in medical surgical units

(Levy et al., 2010). Roughly 24% of patients who develop severe sepsis or septic shock

will do so in a medical-surgical unit (Tazibir, 2012). The SSC strongly encourages early

treatment and sepsis recognition through the guidelines, bundle recommendations, and

early goal-oriented therapy (EGDT). Application of such approaches has led to a

continuous quality improvement in sepsis care and has been associated with a reduction

in mortality rate (Dillinger et al., 2013). The utilization of early goal-oriented therapy

(EGDT) and sepsis bundles as recommended by SSC has demonstrated its effectiveness

through the decrease in mortality rates by about 15%-20% (Rivers & Ahrens, 2008). A

study conducted by Marwick and Davey, (2009) on the utilization of the resuscitation and

sepsis bundles concluded that care bundles can indeed improve the reliability of

evidence-based care and patient outcomes in reference to septic patients. However, as

indicated by the authors, it remains to be seen whether such success can be reproduced in

a non-ICU area such as a medical-surgical unit.

4

Significance of the Problem

Since the initial development of the Surviving Sepsis Campaign guidelines

published in 2004 outlining the management of severe sepsis and septic shock, there has

been an absolute disregard on the management of septic patients in medical surgical

units. The lack of attention to medical surgical units in regard to sepsis has formulated a

problem among healthcare providers including nurses on the proper guidelines to

diagnose and treat septic patients (Nelson, LeMaster, Plost, & Zahner, 2009). Although

in recent years there has been an effort to improve the early recognition and treatment of

septic and severe septic patients in non-ICU, there is still much improvement required.

Currently, no unified approach exists to diagnose and treat sepsis in medical surgical

units. More recently with the utilization of prognostic biomarkers for sepsis

identification, there has been more confusion on proper guidelines to treat sepsis in

medical surgical units (Pierrakos &Vincent, 2010). The SSC encourages the

implementation of early evidence-based therapies, but with ICU as the main target.

“Sepsis screening tools have been developed to monitor ICU patients, and their

implementation has been associated with decreased sepsis-related mortality” (Dellinger et

al., 2013, p. 558). However, there is no mention by the SSC on creating a sepsis bundle

for non-ICU.

Healthcare providers have difficulty diagnosing sepsis not only in the ICU and the

Emergency Department but also in the medical-surgical wards (Dillinger et al., 2013). A

study conducted by Carter (2007) identified that 29% of mortality occurred in septic

patients when a sepsis bundle was implemented within the first 24 hours of admission,

versus 49% mortality when the bundle was implemented after 24 hours or not

5

implemented at all. The same study also indicated the necessity to implement a sepsis

screening tool for early sepsis recognition (Carter, 2007).

Given the indistinctness of sepsis diagnosis, the use of biomarkers has become a

new frontier in identifying sepsis, especially lactic acid, procalcitonin (PCT), and serum

soluble triggering receptor expressed on myeloid cells-1(s-TREM-1). Nevertheless, it

presents a dilemma in that there is not sufficient data to support any particular biomarker

as the gold standard for sepsis identification (Pierce, McCabe, White, & Clancy, 2012).

As a result, clinicians must rely on the sensitivity and specificity of sepsis tools and a

combination of various biomarkers.

Problem Statement and Purpose

The problem is that there are no guidelines or protocols on the identification and

treatment of sepsis on medical surgical units. Healthcare providers have difficulty

diagnosing sepsis not only in the ICU and the emergency department but also in the

medical surgical wards (Dillinger et al., 2013).

The purpose of this pilot study was to develop and implement a severe sepsis

bundle on a medical surgical unit to determine if there was a reduction in hospital length

of stay and transfers to higher level of care. The bundle included a severe sepsis

screening tool already utilized at a community hospital in Miami, Florida, a newly

implemented severe sepsis algorithm for early identification of severe septic patients in a

medical surgical non-ICU and a severe sepsis order set. The rationale of the severe sepsis

screening tool was to identify severe sepsis patients before they decompensated and

became hemodynamically unstable. The purpose of the severe sepsis algorithm was to

guide nurses on the appropriate steps to implement early goal-oriented therapy in the

6

event a patient screened positive for severe sepsis. The algorithm delineated the

implementation of two sepsis biomarkers, procalcitonin, and lactic acid. The idea was

that by having medical surgical nurses adhere to the implementation of the severe sepsis

tool and algorithm, they would be able to identify severe septic patients before requiring

transfer to progressive care units or ICUs and consequently reducing the length of

hospital stay.

Following the SCC 2012 sepsis guidelines, a sepsis bundle (screening tool,

algorithm, order set) was developed specifically for medical surgical non-ICU. The

bundle included an already implemented severe sepsis screening tool, a new algorithm

for nurses to follow whenever a patient screened positive for severe sepsis and a set of

orders based on the patient's condition and algorithm outcome.

It is important to understand that although the majority of severe septic patients

require ICU admission, the assessment of sepsis is not solely the domain of the physician,

critical care nurse, or emergency department nurse but of every nurse involved in the care

of the patient (Nelson et al., 2009). All nursing assessments need to take into account the

signs and symptoms of sepsis. This pilot study evaluated severe sepsis patient outcomes

by monitoring patient transfers to higher level of care and sepsis hospital length of stay.

The already implemented severe sepsis tool utilized at a community hospital in Miami,

Florida consisted of an electronic instrument that reports admitted patient information,

such as the individual's latest vital signs and laboratory work in order to detect sepsis

status. This assessment tool was performed by each nurse every shift on every patient

and was only deferred if a patient screened positive for severe sepsis. The screening tool

screened for systemic inflammatory response syndrome (SIRS), sepsis, and severe sepsis,

7

based on the definitions provided by the SSC in 2012. In order to appraise the

significance of the severe sepsis tool, a retrospective study was conducted evaluating the

sensitivity and specificity of the tool.

Project Objectives

The main focus of this pilot study was to provide awareness of sepsis in the

medical surgical units by not only educating the staff but also providing better tools to

easily identify and treat septic patients in non-ICU areas. Given the growing medical and

economic burden of sepsis, by early identification and treatment of septic patients,

practitioners can ensure adequate prevention and therapy for all septic patients, thus

improving severe sepsis incidences and decreasing transfers to progressive care units and

lengthy hospital stays. This pilot study also provided the foundation of a severe sepsis

algorithm and sepsis bundle specifically designed for medical surgical units that could be

adapted into other units and hospitals. The research objectives were designed to improve

the diagnosis of sepsis and identification of patients at risk for severe sepsis and septic

shock, while empowering nurses to implement early goal directed therapy in an attempt

to decrease hospital length of stay and transfers of septic patients to higher level of care.

The objectives of this project were as follows:

1. Generate administration and management support for the project (Appendix B).

2. Evaluate septic patient’s hospital length of stay and patients transfer to higher

level of care on medical-surgical unit prior to implementation.

3. Develop and implement a severe sepsis bundle (Tool, Algorithm, Order Set) for

medical-surgical units:

8

A. Phase 1: Educate medical surgical nurses on the topic of sepsis (not part of

the measurement of the study).

B. Phase 2: Decrease septic patient’s hospital length of stay and decrease

septic patients transfer to higher level of care after implementation of

bundle.

4. Evaluate the sensitivity and specificity of a severe sepsis tool.

The findings of this study had a significant impact in the areas of nursing education,

nursing practice, nursing research, and healthcare policy within the organization.

Theoretical Framework: Donabedian's Model

Donabedian's structure-process-outcome model has been utilized as a useful

theoretical framework for quality assessment, highlighting the significant relationship

between process and outcome (Donabedian, 1998). Donabedian has been recognized as

the father of quality assessment, and much of his life work was dedicated to formulating

the underpinning for improving the quality of the healthcare system (El Haj, Lamrini, &

Rais, 2013). Although born of Armenian parents, Avedis Donabedian studied medicine

in Beirut. Later as a physician, he developed a profound interest in the quality of health,

which led him to travel to the United States in 1953 in pursuit of a public health degree

(Donabedian, 1998). After receiving his Master of Public Health from Harvard

University, Donabedian assembled an extensive quantity of literature regarding health

services measuring the quality of the care being delivered. He later published his

research findings in 1966 with an article titled “Evaluating the Quality of Medical Care”

(Donabedian, 1998). “Evaluating the Quality of Medical Care” outlines all of

9

Donabedian's methodology regarding the assessment and value of care and emphasizes

the breakthrough of his theoretical framework Structure-Process-Outcome.

The Donabedian model is a theoretical framework for exploring health services

and evaluating care by measuring the quality of the services provided (El Haj et al.,

2013). According to the model, quality can be quantified by three concepts: structure,

process, and outcome (Donabedian, 1988). However, although simple, the model

consists of more complex dimensions. Donabedian included seven different attributes to

subdivide the model to evaluate quality: (a) efficacy: the aptitude of care; (b)

effectiveness: the level to which reasonable improvements in health are achieved; (c)

efficiency: ability to generate cost-effective practices without jeopardizing care; (d)

optimality: a system of checks and balances between costs and benefits; (e) acceptability:

conventionality of patient's preferences regarding care; (f) legitimacy: agreement to social

preferences; and (g) equity: equality in the delivery of care (Donabedian, 1990).

Consequently, these seven pillars of care take into account "patient preferences" as well

as "social preferences" in assessing and assuring quality (Donabedian, 1990).

In order to understand the model, it is important to define quality of care. Many

scholars have attempted to clarify this concept, including Donabedian himself; however,

the Institute of Medicine (2001) provided a widely accepted terminology to define

quality of care: “The degree to which health services for individuals and populations

increase the likelihood of desired health outcomes and are consistent with current

professional knowledge” (p. 2). The model does not only highlight the quality of care but

aims to assess care by evaluating performance of practitioners (Donabedian, 1988).

10

According to Donabedian (1988), performance can be evaluated in two

categories: technical performance and interpersonal performance. Technical performance

refers to the knowledge and judgment necessary to arrive at a specific strategy to

implement an outcome. Interpersonal performance refers to the ability of the practitioner

to conclude a diagnosis with the appropriate method of care (Donabedian, 1988). These

two concepts are important when interpreting Donabedian's structure-process-outcome

model because it provides the foundation for the assessment of quality care. The model

focuses on the importance of quality, suggesting that evaluation of care can be

conceptualized into three dimensions: structure, processes, and outcomes (Donabedian,

1966). The three concepts are built interchangeably: “there must be preexisting

knowledge of the linkage between structure and process, and between process and

outcome, before quality assessment can be undertaken” (Donabedian, 1988, p. 1745).

Donabedian proposed that practitioners could evaluate the level of quality by examining

the formation of the surroundings in which care is being rendered through process

assessment and outcomes measurement.

Structure refers to the environment in which care is delivered, measuring

organizational characteristics such as staffing, practitioner's skills, and financial

resources. Following this approach, high-quality care settings and supporting structures

denote good care. Structure can be viewed as not only the organization configuration and

operation, but also as the policies and procedures implemented to achieve high quality

care. An example of this phenomenon may be the expectation that all staff are aware of

their functions and errands within an organization, thus working together to generate high

quality care (Clearly & O'Kane, 2013). Structure can be subdivided into three concepts:

11

patient characteristics, system characteristics, and providers' characteristics (Donabedian,

1980). Patient characteristics refer to data involving patients, such as demographics,

gender, admitting diagnosis, personal beliefs, and premonitions. System characteristics

refer to resources, accessibility, and design (Donabedian, 1980). Lastly, providers’

characteristics include not only the physician but also the nursing staff rendering the care,

instituting the education, gender, demographics, beliefs, and attitudes of the staff

(Donabedian, 1980).

Process contributes to the interactions between healthcare professionals and their

clients. According to Donabedian (1988), this concept highlights the relationship

between how care has been provided by the practitioner and the responsibilities required

of each patient. Instruments that assess process variables are classified under the

following headings: communication, patient knowledge, performance appraisal, and

quality of care (Donabedian, 2005). The measurement of process is comparable to the

measurement of quality of care because process contains all aspects of healthcare

delivery. Process can delineate the quality of the services delivered; in other words,

whether "good" care has been distributed (Donabedian, 2005).

Outcomes signify the end product of care, such as functional improvement,

revitalization, or survival. Outcomes are usually tangible and accurately calculated

(Donabedian, 1988). This approach defines the effectiveness of the treatment

implemented and the satisfaction of each healthcare consumer. Instruments that assess

outcome variables are classified under the following headings: patients, health providers,

and organization health system (Donabedian, 2005). “If quality-of-care criteria based on

structural, process, or intermediate outcomes are to be credible, it must be demonstrated

12

that variations in the attribute they measure lead to differences in health status outcomes”

(Donabedian, 1988, p. 1744). Outcome can be utilized not only to evaluate expected or

unexpected circumstances between patient-practitioner relationships but also to evaluate

specific organizational benchmarks to evaluate quality of care, such as National Database

of Nursing Quality Indicators (NDNQI). Donabedian's model has been described as a

linear framework model consequently influencing and interlocking the three components

of structure, purpose, and outcome with each other.

Theoretical Framework as It Relates to Clinical Problem

An analysis of the three main concepts of this theoretical framework: structure,

process, and outcomes were essential in identifying and rapidly treating severe septic

patients in medical surgical units. Donabedian's structure-process-outcome model was

the most appropriate theoretical framework for evaluating a severe sepsis tool already

utilized at a community hospital in Miami, Florida and implementing a severe sepsis

algorithm for early identification of severe septic patients in a medical-surgical non-ICU.

The framework intended to reduce patient transfers to higher levels of care and hospital

length of stay (Adaptation of structure, process, and outcome model Donabedian, 1980;

see Figure 1). As previously stated, this model provides the theoretical framework for

assessment of quality care in various situations encountered by today's healthcare

professionals. The framework can be utilized to transform structures and processes

within any healthcare delivery unit such as small clinics, ambulatory care centers, and

rural and urban hospitals. In addition, the Donabedian's model is appropriate for the

organization and development of treating multiple diseases and debilitating conditions

with the aim to improve the quality of disease management such as sepsis (Lawson &

13

Yazdany, 2012). Following Donabedian's framework, identification of sepsis can be

parted into the three components of the theory: structure, process, and outcome

(Donabedian, 1988).

Structure, following Donabedian's model was composed of physical resources and

organizational resources. Physical resources included supplies, adequate staffing, patient

population, physicians' collaboration, and nurses' critical thinking proficiency.

Organizational resources incorporated leadership, transparency, human resources, and

specific organization expectation (Dougherty, 2012).

Process, in reference to this capstone study, included interactions between the

healthcare practitioner (physicians, nurse practitioners, nurses) and the patient population.

Significant components were implemented in the study's process measures, which

included a severe sepsis algorithm with biomarkers procalcitonin and lactic acid,

guidelines and order sets, and a proposed course of action. Process, conceptually was

referred to the service provided leading to a desired outcome. Consequently, the process

of the severe sepsis screening tool was to identify severe septic patients before

decompensating and becoming hemodynamically unstable, while the purpose of the

severe sepsis algorithm was to guide nurses on the appropriate steps to implement early

goal-oriented therapy in the event a patient screened positive for severe sepsis.

Outcomes in this study referred to the method of analyzing the collected data.

The study attempted to evaluate if the implementation of a severe sepsis algorithm

decreased sepsis hospital length of stay and transfers of septic patients to higher level of

care. In addition, the research focused on formulating high-quality clinical results, by

identifying severely septic patients before becoming hemodynamically compromised.

14

Intermediate outcomes were also measured, such as evaluating biochemical markers for

sepsis identification (procalcitonin and lactic acid). According to Donabedian (1988)

“knowledge about the relationship between structure and process (or between structure

and outcome) proceeds from the organizational science” (p. 1745).

Significance of the Study

Healthcare Practice

Early sepsis identification still remains the greatest challenge when managing

patients with sepsis. Following the recommendations of Surviving Sepsis Campaign

(SSC) of 2012, many organizations have strived to implement screening tools and

bundles for the treatment of sepsis and severe sepsis. However, it has not been focused

on medical surgical units. Recent studies demonstrate that roughly 24% of patients who

develop severe sepsis or septic shock will do so in a medical-surgical unit (Tazibir,

2012). The absence of targeted research for severe sepsis in these units suggests the need

for further study.

Improving sepsis awareness in medical surgical units by simultaneously educating

the staff and providing better tools to both identify and treat septic patients are now a

priority. Despite the growing medical and economic burden of sepsis, through early

identification and treatment of septic patients, nurses can ensure adequate prevention and

therapy for all septic patients. Early sepsis identification has demonstrated improvement

in decreasing severe sepsis incidences on intensive care units; therefore, this study can

potentially decrease transfer to higher level of care and lengthy hospital stays. By

creating a severe sepsis algorithm and sepsis bundle specifically designed for medical

surgical units, nurses were able to not only identify septic patients earlier but also

15

provided adequate early goal directed therapy. This model that was piloted in a specific-

medical surgical unit at a community hospital in Miami, Florida may perhaps have the

ability to be adopted into sister units and hospitals. With the creation of such algorithm

and bundle, the researcher attempted to reduce not only the incidences of severe sepsis

patients but also to reduce septic transfer to ICUs and reduce hospital stay, while

assisting physicians and practitioners to better diagnose patients with sepsis.

Additionally, this pilot study added research to the body of knowledge allocated with

sepsis.

Healthcare Outcomes

Sepsis prevalence is increasing not only in the United States but worldwide.

Although multiple speculations exist regarding the increased occurrence of sepsis, either

because of improvement in diagnosis or due to demographic changes in populations, the

main goal remains the same, to improve patient outcomes through early identification and

aggressive treatment. The development of a severe sepsis algorithm and sepsis bundle

specifically designed for medical-surgical units can also improve sepsis mortality and

morbidity rates because therapy will be implemented earlier, potentially improving

patient outcomes. This project impacted healthcare outcomes by reducing patient transfer

to higher level of care and by reducing length of hospital stays; as a result, hospitals will

be able to alleviate the cost per unit and hospital expenses.

Healthcare Delivery

The intent of this capstone project was to significantly impact the healthcare

delivery system in an effort to improve awareness about the subject of sepsis in medical

surgical units not only locally but potentially regionally. Through increased knowledge

16

of sepsis, organizations and practitioners may be able to positively influence the

healthcare delivery pertaining to the topic of sepsis. One of the most important

recommendations of the Surviving Sepsis Campaign (SSC) of 2012 is the utilization of

sepsis tools and bundles for early sepsis identification and early goal-oriented therapy,

advocating healthcare institutions and clinicians to be actively engaged in the sepsis

movement. This capstone project employed a severe sepsis screening tool already

utilized at a community hospital as well as implemented a new severe sepsis algorithm

for early identification of severe septic patients in a medical-surgical non-ICU, in an

effort to reduce patient transfers to higher levels of care and hospital stays. Additionally,

the creation of this algorithm may result in several other positive outcomes such as

reducing hospitals expenses, decreasing sepsis readmission rates, and improving sepsis

mortality and morbidity. Sepsis complications in admitted patients may also be reduced

by actively identifying septic patients and by providing adequate aggressive therapy

modalities.

Another important aspect to consider when discussing the topic of the healthcare

delivery and sepsis is the subject of core measures. Core measures are assessment

formats that the government utilizes to improve the quality of hospital care while

reducing costs; the four major core measures are: acute myocardial infarction, heart

failure, pneumonia, and surgical care (Center for Medicare & Medicaid Services, 2014).

However, given the reality that the four major types of core measures (especially

pneumonia and surgical site) can occur simultaneously with sepsis, the Center for

Medicare & Medicaid Services (CMS) has recently contemplated adding sepsis as a core

measure (Center for Medicare & Medicaid Services, 2014). This notion has driven

17

organizations and practitioners to be more assertive and vigorous in the management of

septic patients, which has compelled them to create sepsis bundles and algorithms to

improve quality outcomes, such as the one proposed in this capstone project.

Healthcare Policy

Healthcare policies are extremely important to our healthcare model, mostly

because it deals with a section of legislation that focuses on the decisions, plans, and

actions to attain specific goals within our population (WHO, 2011). Through the

utilization of research, evidence-based practice, and performance improvement,

policymakers are capable of improving healthcare policy relevant outcomes. According

to the World Health Organization (2011), an unambiguous health policy is one that can

achieve multiple objectives, such as delineating the future, highlighting specific groups’

expectations, and building trust and informing its end users.

Sepsis, much like other healthcare issues, is becoming more noticeable among

policymakers. This visibility could be because of its increased prevalence and high

mortality rate; now more than ever is the perfect time to propel improvement through

policy to improve the nation’s sepsis outcomes. In 2013, New York became the first

state in the nation to mandate every hospital to adopt evidence-based protocols for the

early diagnosis and treatment of sepsis (Cuomo, 2013). Following the SSC

recommendations and the endorsement of the National Quality Forum, New York

hospitals are now required to utilize screening tools for early recognition of patients with

sepsis, severe sepsis, and septic shock, in addition to early goal-oriented therapy for the

management of septic patients. Through the implementation of a severe sepsis algorithm

for early identification of severe septic patients in a medical surgical non-ICU, this

18

capstone project has the potential to generate policy changes beyond the local municipal

government; it may be capable of radiating its impact to state legislation such as it had

occurred in New York.

Summary

Sepsis is a life-threatening condition that affects not only critically ill patients in

ICUs but also in medical surgical units. Early identification and treatment of septic

patients has demonstrated significant decrease in mortality (Rivers & Ahrens, 2008).

Severe sepsis has largely been studied in the intensive care unit. Limited research is

available regarding sepsis identification and treatment in medical-surgical units. This

capstone project focused on the implementation and evaluation of a severe sepsis

algorithm in a medical-surgical unit of a large community hospital in Miami, Florida.

19

Chapter 2: Literature Review

This chapter will provide a comprehensive review of the literature that is relevant

to sepsis identification and management. The literature review contains evaluation of

epidemiology studies in the area of sepsis, assessment of the implemented management

guidelines, identification strategies for successful implementation of the Surviving Sepsis

Campaign recommendations, and systematic review of the usage of biomarkers in the

treatment of sepsis. A comprehensive literature review was performed utilizing articles

obtained through several literature searches from the following databases: Cumulative

Index for Nursing and Allied Health Literature (CINAHL), PubMed; Cochrane Library;

MEDLINE; Science Direct, Up-to-date, Baptist Health Library System, NOVA Library

System, and MD Consults. The searches were limited to studies primarily conducted

within the last 5 years, 2008-2013 in the English language; however, there were four

considerably important studies from earlier years (one from 2001 and three from 2007).

The pediatric population was excluded given differences in criteria on sepsis and

diminutive resemblance to the adult population. The search criteria utilized specific

keywords: sepsis, severe sepsis, sepsis bundles, surviving sepsis campaign, systemic

inflammatory response syndrome (SIRS), epidemiology, sepsis biomarkers,

procalcitonin, septic shock, and protocol. All articles were sorted into multiple categories

including: epidemiology, treatment recommendations and modalities, implementation

strategies, biomarkers for sepsis identification, and procalcitonin biomarker as a de-

escalation of antibiotic therapy. A total of 68 articles were reviewed, with 29 articles

20

included in this review. Articles that were excluded from this review were based on the

applicability to the project.

Epidemiology

Angus et al.’s (2001) study concluded that sepsis is the primary cause of death in

the non-coronary intensive care units and one of the most significant challenges in critical

care. An extensive review of more than 6,600,000 cases was conducted, yielding

staggering results. The authors concluded that 192,980 cases were identified positive for

severe sepsis following the International Classification of Diseases, Ninth Revision,

Clinical Modification classification. Angus et al. (2001) also concluded that out of

751,000 cases (3.0 cases per 1,000 population and 2.26 cases per 100 hospital

discharges), 383,000 (51.1%) received intensive care treatment and an additional 130,000

(17.3%) were ventilated in an intermediate care unit or cared for in a coronary care unit.

Angus et al. (2001) revealed prevalence increased >100-fold with age (0.2/1,000 in

children to 26.2/1,000 in those >85 yrs old). Mortality was approximately 30%, with an

increase in predominance with advanced age. The study not only revealed the

epidemiology of sepsis but accentuated the average cost sepsis per case. The average

costs per sepsis case were $22,100, with annual total costs of $16.7 billion nationally

(Angus et al., 2001).

A more recent study conducted by Hall et al. (2011) in collaboration with the

National Center for Health Statistics (NCHS) evaluated the inpatient care for sepsis

during 2008 based on the data from the National Hospital Discharge Survey in the United

States. The study concluded similar results to Angus et al.’s (2001) study. The number

and rate per 10,000 hospitalizations for sepsis or severe sepsis has more than doubled

21

from 2000 to 2008, with incidence increasing approximately 5-10 % each year. During

the same period, the hospitalization rate for those with septicemia or sepsis as a principal

or as a secondary diagnosis increased by approximately 70% from 22.1 to 37.7 per

10,000 population. The study highlighted that, as Angus et al. pointed out, an aging

population with more chronic illnesses has a higher possibility of acquiring sepsis. Other

possible attributing factors for sepsis could be greater use of invasive procedures,

immunosuppressive drugs, chemotherapy, and transplantation, and increasing microbial

resistance to antibiotics. Hall et al. (2011) concluded that the nationwide inpatient annual

costs for treating those hospitalized for septicemia were estimated to be $14.6 billion.

Lastly, the study also concluded that those patients who did survive severe sepsis were

more likely to have negative long-term effects on their health and mental being.

Wang, Shapiro, Angus, and Yealy (2007) conducted an emergency department

(ED) multisite study collecting data from 2001-2004 to evaluate the nation's average of

the number, timing, ED length of stay, and case distribution of patients presenting to the

ED with suspected severe sepsis. The study concluded that during the 4-year span,

approximately 2.3 million (571,000 annually, 0.69%; 95% confidence interval [CI], 0.61–

0.77%) patients were suspected of acquiring severe sepsis; with the mean length of stay

for suspected severe sepsis of about 4.7 hour (95% CI, 4.3–5.1 hrs). The study again

validated that sepsis prevalence and mortality increases with age (Wang et al., 2007).

Consistent with Wang's findings, Elixhauser, Friedman, and Stranges (2011) suggested in

their study that the highest rates of sepsis occurred among the elderly. The authors aimed

to provide data collected from the Healthcare Cost and Utilization Project (HCUP)

Nationwide, Inpatient Sample (NIS) on hospital stays for septicemia from 2000 and 2009.

22

In the study, data between the two years were evaluated demonstrating that

approximately 1,920 hospital stays occurred for septicemia per 100,000 people ages 65-

84 and 4,020 septicemia stays per 100,000 people ages 85 and older during 2009

(Elixhauser et al., 2011). In the same study, the authors also concluded that sepsis

identification is important to decrease mortality and morbidity and is extremely

expensive to treat. Between 2000 and 2009, the average length of stay and average

inflation-adjusted cost of a stay with a principal diagnosis of sepsis syndrome both

increased (from 8.2 days to 8.8 days and from $12,800 to $18,500, respectively), as

expected this directly affected the cost per day from $1,700 to $2,300 (Elixhauser et al.,

2011). Cost per stay and cost per day for hospitalizations with a secondary diagnosis of

sepsis syndrome also increased between the two years (from $30,900 to $33,900 and

from $2,100 to $2,300, respectively). It has been estimated that approximately $17 billion

is spent annually in the United States for the treatment of sepsis and sepsis syndrome

(Elixhauser et al., 2011).

Treatment Recommendations and Modalities

In an approach to improve the increasing rate and complications of sepsis and

severe sepsis, healthcare providers have formulated different methods to not only rapidly

identify sepsis but also treat it. As a result, the Surviving Sepsis Campaign (SSC) was

launched in 2004, and a set of guidelines were created for sepsis management (Tromp et

al., 2011). Within the guidelines recommended by the SSC, the resuscitation bundle and

the management bundle are among the most important components for the treatment of

septic patients (Dillinger et al., 2013). The SSC strongly encourages early treatment and

sepsis recognition through the guidelines, bundle recommendations, and early goal-

23

oriented therapy (EGDT). Application of such approaches has led to a continuous quality

improvement in sepsis care and has been associated with a reduction in mortality rate

(Dillinger et al., 2013). The utilization of early goal-oriented therapy (EGDT) and sepsis

bundles as recommended by SSC has demonstrated its effectiveness through the decrease

in mortality rates by about 15-20% (Rivers & Ahrens, 2008). A study conducted by

Marwick and Davey (2009) on the utilization of the resuscitation and sepsis bundles

concluded that care bundles can indeed improve the reliability of evidence-based care and

patient outcomes in reference to septic patients. However, as indicated by the authors, it

remains to be seen whether such success can be reproduced in a non-ICU area such as a

medical-surgical unit.

The SSC of 2012 followed the same grading system as its predecessor, Grading

of Recommendations Assessment, Development and Evaluation (GRADE) system to

evaluate and properly assess the quality of each recommendation. Letters A (High) to D

(Very Low) were appointed to each recommendation on the basis available research and

a strength scale of 1 (strong) and 2 (weak) was also utilized (Dellinger, et al., 2013).

According to the SSC, the initial resuscitation and management of severe sepsis is aimed

to identify and aggressively treat patients with tissue-hypoperfusion and achieve

hemostasis. Resuscitation should begin within the first 6 hours of identifying sepsis.

Clinicians must be diligent in treating septic patients with attention to CVP monitoring 8-

12 mm Hg, MAP >or =65 mm Hg, urine output > or equal 0.5 ml/kg/hr and Scvo2 of 70

%. Lactic acid should be monitored as well and the hypoperfusion protocol started if

lactic acid is found to be greater than 4 mmol/l (Dellinger et al., 2013).

24

Accurately identifying sepsis at an early stage can potentially decrease mortality

by 5-10%, consequently accentuating the importance of early detection and prompting

therapeutic treatment (Singer, 2013). Connecting early sepsis recognition with pathogen

identification allows for practitioners to effectively administer proper antibiotic therapy

and treatment more rapidly, hence decreasing mortality and morbidity (Singer, 2013).

The SSC recommends obtaining appropriate blood cultures before the administration of

antibiotic therapy if it occurs within 45 minutes of sepsis identification. It is

recommended for the utilization of broad-spectrum antibiotic therapy that will cover

gram positives, gram negatives, and atypical pathogens. De-escalation of antimicrobial

therapy plays an important role in sepsis management; antibiotics should be discontinued

once the patient is identified as non-toxic. Fluid resuscitation plays an important role in

properly treating septic patients. The SSC recommends the utilization of crystalloids

30ml/kg as the initial fluid challenge in patients with sepsis, with the utilization of

vasopressors if fails (Dellinger et al., 2013). Lastly, the SSC recommends the general

assessment of the patient to properly identify and highlight the source of infection and

provide adequate treatment; such as abscesses, wounds, infected catheters (whether Foley

or central line), implantable devices, etc.

A retrospective cohort study conducted by Rhode et al. (2013) to determine the

rates of infection and organ dysfunction in patients with sepsis admitted to non-ICU

wards concluded that severe sepsis was not only found on different non-ICU but was

poorly documented. More than 20,000 patients were screened for severe sepsis;

approximately 14% screened positive for severe sepsis, which included a mean age of 63

years. Cardiovascular and renal failure were found to be the most common type of organ

25

dysfunction in about 66% of patients (Rohde et al., 2013). A similar study conducted by

Carter (2007) aimed to identify how a critical care team at a metropolitan hospital

implemented a severe sepsis bundle in the ED and medical surgical units. The study

concluded that 29% of mortality occurred in septic patients when a sepsis bundle was

implemented within the first 24 hours of admission, versus a 49% mortality when the

bundle was implemented after 24 hours or not implemented at all. The same study also

indicated the necessity to implement a sepsis screening tool for early sepsis recognition

as recommended by the SSC (Carter, 2007).

Implementation Strategies

It is clear that sepsis is not only a growing pandemic but also a very deadly and

costly situation for patients as well as institutions. Prevention is the best defense strategy

available; however, if severe sepsis develops early, goal-directed therapy (EGDT) has

been shown to decrease prevalence and mortality (Dellinger et al., 2013). Nurses play a

vital role in the prevention, implementation strategies, and treatment of septic patients in

the hospital setting (Tazibir, 2012). Nurses are essential in identifying septic patients

through the utilization of simple tools and robust innovated bundles; nurses have the

potential to intervene early to prevent sepsis mortality and morbidity (McClelland &

Moxon, 2014).

However, implementing all the recommendations proposed by the SSC have not

been an easy task, either because of insufficient staff education or compliance

disengagement; there have been multiple methods utilized to integrate these

recommendations into practice. There have been multiple studies and publications

supporting the different methods utilized to implement these guidelines. Smith, Rice, and

26

Winterbottom (2012) published an article studying the significant value of nurses in

improving adherence to the implementation of EGDT. The authors suggested that

educating staff about sepsis syndrome and sepsis management through “translation of

best practice” is the best approach in improving outcomes (Smith et al., 2012).

Burney et al. (2012) conducted a similar study identifying barriers to the

implementation of a sepsis protocol approach in a New York metropolitan hospital as

suggested by the SSC (2008). The authors conducted an online survey to nurses and

physicians investigating baseline knowledge and confidence in self identifying sepsis,

current practice treatment, management of sepsis, perceived barriers in treating and

identifying sepsis, and suggestions. A sample of 101 staff were surveyed; the study

identified three major barriers with resuscitation protocols for sepsis: first delay in

diagnosing by physicians, second delay in availability of transfers, and third inability to

accurately know how to treat septic patients. The participants, however, agreed that a

sepsis protocol would not only be helpful but necessary (Burney et al., 2012).

A similar study conducted by Durthaler, Ernst, and Johnston (2009) aimed to

evaluate the generalized knowledge regarding sepsis and sepsis management among

different U.S. hospitals. In this study, the authors conducted a series of questionnaires to

nurse managers that were members of the Association of Critical Care focusing on the

management of severe sepsis including policies and treatment plans. More than 400

hospitals participated in the study, concluding that of the 17 SSC treatment guidelines,

hospitals most frequently adhere to blood culture handling, utilization of broad spectrum

antibiotics, and deep venous thrombosis prophylaxis (Durthaler et al., 2009).

27

Giuliano, Lecardo, and Staul (2011) conducted a study of the evaluation and

adherence to the SSC guidelines of a clinical decision support system such as Protocol

Watch to improve septic patients care and outcomes in the Intensive Care Unit. The

study resulted an improvement in compliance with the resuscitation bundle (p = .01) and

decrease time to administer antibiotics (p = .006), thus potentially contributing to a

reduction in morbidity and mortality in ICU septic patients. However, no significant

changes occurred for compliance with the management bundle or resuscitation bundle.

Subsequently, Thiel et al., (2009) conducted a retrospective study in a St. Louis

Metropolitan Hospital to evaluate the hospital-wide impact of a standardized order set for

the management of severe sepsis. The study concluded that after examining 400 patients

the management of severe sepsis was associated with superior fluid administration (p =

.04), enhanced antibiotic therapy (p = .01), decreased incidence of organ failure, and

improved survival (p < .01) (Thiel et al., 2009). Overall, nurses play a pivotal role in

sepsis recommendations, implementation strategies, and the utilization of therapies based

on the latest scientific evidence. Sepsis protocols have shown not only to decrease

mortality but also have resulted in considerable savings for institutions and third party

payers. Broader implementation of sepsis treatment protocols represents a potential

means for enhancing resource use while controlling costs (Dellinger et al., 2013).

Biomarkers for Sepsis Identification

Biomarkers have been suggested as ways to identify sepsis and severe sepsis for

early recognition and prompt antibiotic therapy. A biomarker is defined as “a

characteristic that is objectively measured as an indicator of normal biological processes,

pathogenic processes, or pharmacological responses to a therapeutic intervention”

28

(Sankar & Webster, 2013, pp. 269). An ideal sepsis biomarker is one that not only is

easy to read and rapidly available but also one that has a high specificity and sensitivity

in diagnosing sepsis (Sankar & Webster, 2013). However, finding one specific

biomarker with all of the above characteristics is nearly impossible. Some sepsis

biomarkers have shown promising results in the fight against sepsis such as procalcitonin

(PCT) and soluble triggering receptor expressed on myeloid cells-1 (s-TREM-1),

especially in non-ICU patients, thus improving the manner in which sepsis is managed

(Singer, 2012).

A biomarker should reflect the pathogenic process of the disease and respond to

intervention. The use of biomarkers has become a new frontier in identifying sepsis,

especially lactic acid, PCT, and s-TREM-1. Nevertheless, it presents a dilemma in that

there is not sufficient data to support any particular biomarker as the gold standard for

sepsis identification (Pierce et al., 2012). Recent literature suggests the utilization of

multiple biomarkers for sepsis diagnosis and treatment, thus preventing further

deterioration. A review of all readily available biomarkers conducted by Sanker and

Webster (2013) concluded that more research needs to be conducted to clearly identify a

gold standard for sepsis identification. More than 175 inflammatory biomarkers were

described and reviewed, ranging from cultures, CRP, PCT, and genetic biomarker,

without clearly optimizing one over other.

Gamez-Diaz et al. (2011) conducted a study to evaluate the diagnostic accuracy of

high-mobility group box-1 protein (HMGB-1), cluster differentiation-64 (CD64), and

serum levels of soluble triggering receptor (s-TREM-1) as sepsis biomarkers evaluating

the sensitivity and specificity in a busy emergency department. A total of 631 patients

29

were included in the study, 66% (n = 416) were diagnosed with sepsis syndrome based

on clinical presentation and those biomarkers were drawn. The study concluded that

HMGB-1 has a sensitivity of 57.5% and specificity of 57.8%, CD64 has a sensitivity of

65.8% and specificity of 64.6% and s-TREM-1 has a sensitivity of 60% and specificity of

59.2%. The study also concluded that in this cohort of patients, none of these three

biomarkers demonstrated significant sensitivity or specificity to be utilized as a

diagnostic tool (Gamez-Diaz et al., 2011). A similar study conducted by Zhang, She,

Feng, Jia, & Xie (2011) yielded similar results. Zhang, She, Feng, Jia, and Xie (2011)

studied the utilization of serum levels of s-TREM-1 for the diagnosis and severity

assessment of septic patients and compared the values to C-reactive protein (CRP) and

procalcitonin (PCT). The study concluded that s-TREM-1 levels reflected the severity of

sepsis more accurately than those of CRP and PCT; however, the sensitivity for s-TREM-

1 was 59.5% while specificity was 93.3%.

Gibot et al. (2012) conducted a prospective study utilizing 79 critically ill patients

to determine the usefulness of combination biomarkers to diagnose sepsis. Plasma

concentration of s-TREM-1, PCT, CD64, and high affinity immunoglobin-Fc fragment

receptor (FcyRI) were measured, and a bioscore combining these biomarkers was

created. The study concluded that sTREM-1, PCT, and CD4 were all independent

predictors of an infectious process with a p-value <.001. The sensitivity and specificity

of these three biomarkers demonstrated an improved clinical performance; PCT

sensitivity 83.1% and specificity 84.9%, sTREM-1 sensitivity 53.2 and specificity 86.3

and CD64 sensitivity 84.4 and specificity of 95.2 (Gibot et al., 2012). A similar study

comparing eosinopenia as a marker of sepsis to other sepsis biomarkers concluded that

30

eosinopenia is a very sensitive serologic biomarker for identifying critically ill septic

patients, but not very specific. However, the study revealed that eosinopenia was

considerably inferior to PCT and CRP as a diagnostic biomarker (Shaaban, Daniel,

Sison, Slim, & Perez, 2010). Eosinepenia as a diagnostic sepsis biomarker was compared

to PCT and CRP levels. A total of 68 patients were enrolled in the study, and the results

yielded a CRP sensitivity of 94% and specificity of 84%, PCT sensitivity 84% and

specificity of 92%, and eosinophil sensitivity of 81% and specificity of 65% (Shaaban et

al., 2010).

An observational, prospective study was conducted on 50 critically septic patients

to establish the prognostic and discriminative value of pro-arterial natriuretic peptide

(pro-ANP) as a sepsis biomarker and compare to those of PCT (Lipinska-Gediga,

Mierzchala, & Durek, 2012). The study concluded that pro-ANP was significantly higher

in non-survivor patients (p < .05), pro ANP in contrast to PCT failed to possess

significant diagnostic and discriminative value (Lipinska-Gediga et al., 2012). A similar

prospective study conducted by Hoeboer et al. (2012) on 101 critically ill patients

comparing the role of old and new biomarkers in predicting absence or presence of

infection concluded that PCT was the single best predictor for low risk of infection and

CRP and lactic acid combined were the best predictor for high risk infection follow by

PCT. According to the study, only peak PCT (cutoff 0.65 ng/ml) was a true predictor for

all endpoints studied (Hoeboer et al., 2012). Another prospective study to evaluate the

diagnosis and prognosis values of PCT, interleukin-6 (IL-6), CRP, erythrocyte

sedimentation rate, and white blood counts in septic patients yielded similar results

(Jekarl et al., 2013). Approximately 177 patients were included in the study, of which 78

31

were diagnosed with sepsis criteria. The study concluded that PCT was superior in

diagnosing sepsis in comparison to the other biomarkers with a 74.4% and 93.7%

sensitivity and 86.7% and 75.2% specificity. However IL-6 demonstrated superior

kinetics for monitoring the effectiveness of antibiotic therapy (Jekarl et al., 2013).

In contrast, Tang, Eslick, Craig, and McLean (2007) conducted a systematic

review and meta-analysis study to estimate the diagnostic accuracy of procalcitonin in

sepsis diagnosis in critically ill patients. A total of 18 studies were investigated,

concluding that the diagnostic performance of procalcitonin was low, with mean values

of sensitivity 71% (95% CI 67-76) and specificity 71% (95% CI 67-76) and an area under

the ROC curve of 0.78 (95% CI 0.73-0.83). The authors concluded that procalcitonin

cannot reliably differentiate sepsis from other non-infectious causes of systemic

inflammatory response syndrome in critically ill adult patients (Tang, Eslick, Craig, &

McLean, 2007). However, Wacker, Prkno, Brunkharst, and Schlattmann (2013)

conducted a systematic review and meta-analysis on the accuracy and clinical value of

procalcitonin as a sepsis biomarker in the critically ill population. A total of 3,487

records were obtained, of which 30 were utilized based on the inclusion criteria,

accounting for a total of 3,244 patients. Bivariate analysis was calculated, resulting in a

mean sensitivity of 0.77% (95% Cl 0.072-0.81) and specificity of 0.79 (95% Cl 0.74-

0.84) with a ROC curve of 0.85 (95% Cl 0.81-0.88). The study concluded that

procalcitonin is indeed a helpful sepsis biomarker for early identification of septic

patients.

32

Procalcitonin Biomarker as a De-Escalation of Antibiotic Therapy

Procalcitonin is a peptide that usually rises in blood because of inflammatory

reactions and presence of infection. PCT is currently being utilized in the medical

practice as stewardship for antibiotic de-escalation, detection of new infections, and

exclusion of sepsis diagnosis (Wolff & Bouadma, 2010). The SSC recommended that

antibiotic therapy should begin within the first hour of identifying severe sepsis. The

SSC also recommended the utilization of low procalcitonin levels or similar biomarkers

to assist the clinician in the discontinuation of empiric antibiotics in patients who initially

appeared septic but have no subsequent evidence of infection (grade 2C) (Dellinger et al.,

2013).

Oh et al. (2009) conducted a prospective study to evaluate the usefulness of

utilizing procalcitonin fast kit as a guideline for commencing antibiotic therapy on septic

patient in an emergency department. A total of 80 patients were included in the study; 33

of them were identified as severe sepsis and PCT levels were drawn. The study

concluded that PCT is a useful method for detecting severe sepsis in the ED with a

sensitivity of 93.94% and specificity of 87.23% with an area under the curve of 0.916

(Oh et al., 2009).

Kopterides, Siempos, Tsangaris, Tsantes, and Armaganidis (2010) conducted a

review of all randomized controlled study to evaluate the effectiveness of a procalcitonin-

guided algorithm for de-escalation of antibiotics in septic patients in an intensive care

unit. Seven randomized controlled studies were included for a total of 1,131 number of

participants. The implementation of procalcitonin-guided algorithm demonstrated a

decreased in the utilization of antibiotic therapy by approximately 2 days and a total

33

duration of antibiotic therapy by 4 days (Kopterides et al., 2010). A similar retrospective

study conducted by Hohn et al. (2013) aimed to analyze the effects of antibiotic therapy

in severe septic patients after the implementation of a PCT protocol. A retrospective ICU

search was conducted from 2005-2009 on patients with sepsis syndrome criteria who

were treated accordingly to PCT guided algorithm. A total of 141 patients were included,

and primary outcome parameters were calculated utilizing ANCOVA. The study

concluded that the implementation of a PCT-algorithm protocol was in fact associated

with a reduction of antibiotic therapy without compromising clinical outcomes. In 2005,

there was a reduction of 1.0 day per year from 14.3 _+ 1.2 to 9.0_+1.7 days in 2009 (p =

0.02). There was also a 28-day mortality reduction in average year by 35.1% and a mean

cost reduction of -14.3 Euros (Hohn et al., 2013).

Summary

Early recognition and treatment of sepsis still remains the priority among

clinicians to decrease morbidity and mortality in our healthcare system. The

recommendations proposed by the SSC provide practitioners with the foundations for

treatment and implementation when dealing with septic patients. The review of the

literature revealed that early identification and aggressive treatment is necessary to

improve patients’ outcome; with the usage of identification tools and biomarkers,

clinicians will be able to provide a better diagnosis and rapid treatment for the

management of sepsis. However, further research is necessary to accurately promote a

specific biomarker for the management of sepsis; thus, clinicians should not rely solely

on serum tests but on the overall clinical evaluation.

34

Chapter 3: Methods

The purpose of this chapter is to outline the project design, methodology, and a

discussion of the potential outcome of the project. In addition, a review of the methods

and timeline for the clinical practice project will be highlighted. Since the initial

development of the Surviving Sepsis Campaign guidelines published in 2004 outlining

the importance of early identification and management of severe sepsis, there has been an

absolute disregard of the management of septic patients in medical-surgical units,

formulating a problem among healthcare providers including nurses on the proper steps

to diagnose and treat septic patients. The purpose of this pilot study was to employ a

severe sepsis screening tool (Appendix C) already utilized at a community hospital in

Miami, Florida, and newly implement a severe sepsis algorithm (Appendix D) for early

identification of severe septic patients in a medical-surgical non-ICU and severe sepsis

order set (Appendix E), thus potentially reducing patient transfer to higher levels of care

and hospital length of stays.

Project Design

This study used a quasi-experimental, non-randomized one-group pretest-posttest

design. One group pretest-posttest design is a good alternative when intervention impact

is expected to be dramatic and other potential causes have little credibility. This type of

design refers to a study where one group of participants is pretested on the dependent

variable and subsequently post-tested after the treatment condition has been implemented

(Polit & Beck, 2012). In this design, the outcome is representative of the variation

35

between the pretest and posttest data. Still, one group pre test-posttest design does not

control for potentially confounding extraneous variables (history, maturation, testing,

instrumentation, etc.); therefore, it is difficult to identify the effect of the treatment

condition (Polit & Beck, 2012). A retrospective comparison was utilized, accounting for

similar patient groups. Although quasi-experimental design lacks randomization, it is an

exceptionally practical design. Clinical environments provide researchers with

systematic complex situations when conducting a true experimental test. Additionally,

this same approach can sometimes produce research control when full experimental

inflexibility is not possible (Polit & Beck, 2012). Because of the absence of

randomization, occasionally, quasi-experimental design is likely to be accepted by

broader group of participants then true experimental design. However, one major setback

of quasi-experimental design is the lack of internal validity, thus making the cause-effect

relationship less convincing (Polit & Beck, 2012). Frequency matching was utilized to

formulate homogeneous comparable groups. Also, a retrospective review of the data

based on an information technology report was obtained before the implementation of the

algorithm on patients with an ICD-9 diagnosis of sepsis criteria.

Setting

This pilot study was implemented in a 50-bed medical- surgical unit at a

Southeastern Community Hospital. The organization is composed of 680-bed facility

located in the Kendall area of South Miami-Dade County. Founded in 1960, the hospital

offers an extensive variety of medical, surgical, and technological services. Today, the

hospital provides care to a multicultural diverse population including international

patients. Approximately 32,000 clients are hospitalized at this facility yearly, including

36

an estimated 78,000 emergency care visits. It is estimated that approximately 20% of the

admission rate is admitted to the Intensive Care Unit (ICU), accounting for an elevated

number of septic patients (Baptist Hospital of Miami, 2014). It was determined that 4

Tower (a medical-surgical unit) had the highest sepsis length of stay among all other

hospital units; therefore, this pilot study was implemented in the selected unit. Also,

because of the limited amount of information that exists about sepsis in medical surgical

units, the attention of this project was concentrated towards the medical-surgical wards.

Inclusion/Exclusion Criteria

Inclusion Criteria: Pre-implementation of bundle: Patients meeting criteria for

sepsis, severe sepsis, or septic shock based on ICD-9 coding over a period of 3 months

before the implementation of the algorithm were included. Post-implementation of

bundle: All adult patients admitted to the designated medical-surgical unit who screened

positive for sepsis, severe sepsis, or septic shock based on the definition criteria provided

by the Surviving Sepsis Campaign (SSC) and international guidelines over a 3-month

period.

Exclusion Criteria: All patients admitted to pilot medical-surgical unit who did

not qualify with the inclusion criteria were re-screened every 12 hours for evaluation of

sepsis. Patients who were discharged or transferred from the designated medical-surgical

unit without meeting the sepsis criteria were excluded from the study. Comfort measures

only patients and hospice patients were not screened for sepsis and therefore were

excluded from the study. The attending physician had the option to exclude his or her

patient from the study.

37

Recruitment: The study participants included all patients who met the inclusion

criteria. Information technology (IT) data solely contained de-identifiable information

and data based on data points.

Ethical Considerations and Confidentiality

Waiver of Informed Consent

Institutional Review Board (IRB) approval for conducting this study was obtained

from both Nova Southeastern University and the organization where the study was

conducted. The researcher obeyed all policies and procedures related to the protection of

human subjects. Patients were eligible for participation regardless of research objectives

due to its elements as standard of practice. The severe sepsis screening tool and severe

sepsis order set (Appendix E) were standards of care. The severe sepsis algorithm was

implemented as a pilot study proposal and did not alter the standard of care for any septic

patients.

Pre Implementation of bundle: The patients included in the study were already

discharged from the hospital. The patients’ care was not altered. No personal identifiers

were collected during the collection period. The generated report of patients retrieved

from information technology (IT), and the Excel spreadsheet of admission numbers was

password protected on the organization computer system. Once necessary data points

were collected, the password protected list of the patients generated by IT was destroyed.

Identifying and contacting the patients, although not impossible, would not be feasible for

a review of medical records for information that would not change the care they would

have already received.

38

Post-implementation of bundle: IT generated reports and data collection was

utilized de-identified information. Selected data points relevant to the study were

obtained: transfers to higher level of care and hospital length of stay.

Risk and Benefits

De-identified data were electronically kept in a password-protected database on

an organization computer, accessible to the research personnel only. The generated

report of patients retrieved from IT was password protected on the organization computer

system and was kept separate from the data. Once necessary data points were collected,

the password-protected list of the patients generated by IT was destroyed. Although the

potential participants of this study received no direct benefit, the benefit of identifying

patients on a medical-surgical unit at risk for sepsis could result in improved patient

outcomes. These benefits outweighed the risk of a breach in privacy and confidentiality.

The risk of a breach of privacy and confidentiality was minimized by using de-identified

information for the data collection.

Goals and Objectives

The main focus of this pilot study was to provide awareness of sepsis in the

medical-surgical units by not only educating the staff but also providing better tools to

easily identify and treat septic patients in non-ICU areas. Given the growing medical and

economic burden of sepsis, by early identification and treatment of septic patients,

practitioners can ensure adequate prevention and therapy for all septic patients, thus

improving severe sepsis incidences and decreasing transfers to progressive care units and

lengthy hospital stays. This pilot study provided the foundation of a severe sepsis

algorithm specifically designed for medical surgical units that can be adopted into sister

39

units and hospitals. With the creation of such algorithm, an attempt was made to reduce

not only the incidences of severe sepsis with early identification and treatment but also

reduce transfers to ICUs and reduce hospital stay while assisting physicians and

practitioners to better diagnose patients with sepsis.

Project Phases

1. Generate administration and management support for the project (Appendix B).

Administration and management support were obtained by acquiring a letter of

support from the organization’s administration and management department.

Consent to implement the project on the selected medical-surgical unit was

validated with a letter from the nurse manager.

2. Evaluate septic patients’ hospital length of stay and patients transfers to higher

level of care on medical-surgical unit prior to implementation. Pre-

Implementation: Following IRB approval, a request was made to information

technology (IT) for two reports. The first report contained de-identified list of all

patients admitted to the designated medical-surgical unit over a 3-month period

and who presented with a diagnosis of systemic inflammatory response syndrome

(SIRS), sepsis, severe sepsis (including urosepsis), and/or septic shock based on

ICD-9 codes. The second report contained de-identified list all patients admitted

to the designated medical-surgical unit over a 3-month period and who presented

with a diagnosis other than systemic inflammatory response syndrome (SIRS),

sepsis, severe sepsis (including urosepsis), and/or septic shock based on ICD-9

codes. Both reports were extracted from Data Diver, a specific program utilized

within the organization to extract relevant data from a variety of sources.

40

Collected information was saved on an organization's computer and was password

protected. This information was accessible only to the research personnel. De-

identified data was obtained from those reports specifically pertaining to: transfers

to higher level of care and length of stay. Average length of stay and transfers to

higher level of care of septic were calculated.

3. Develop and implement a severe sepsis bundle (tool, algorithm, order set) for

medical surgical units:

A. Phase 1: Educate medical surgical nurses on the topic of sepsis bundle.

Education component: Announcements via flyers (Appendix F) were

provided before the commencement of the pilot study on the selected

medical surgical unit. Education was provided to all registered nurses

employed at the designated medical surgical unit, unit based ARNP, code

rescue team ARNPs, and any voluntary physician regarding sepsis on the

purpose of the pilot study, the significance of the tool and implementation

of the algorithm via face to face encounter and PowerPoint presentation.

The PowerPoint presentation was formulated by the author of this

capstone project following the SSC recommendations. However, the

educational component of this study was not measured as a data point;

therefore, the educational component was not part of the research.

Intervention: Following IRB approval all participating staff (registered

nurses, voluntary hospitalist physicians, and unit-based and code rescue

ARNPs) received a 30-45 minute educational program during staff

meetings, including a PowerPoint presentation and education flyers

41

regarding the study by the principal investigator. The PowerPoint

presentation included the study purpose, objectives, background

information, and education on the proper utilization of the severe sepsis

tool. The presentation also included the proper utilization of the severe

sepsis algorithm.

B. Phase 2: Decrease septic patients’ hospital length of stay and decrease

septic patients’ transfers to higher level of care after implementation of

bundle. Utilizing the severe sepsis tool already in place in the

organization, nurses were required to screen all patients admitted to the

designated medical surgical unit. Bedside nurses followed the severe

sepsis algorithm for those patients who screened positive for severe sepsis.

Attending physicians were contacted in accordance to the severe sepsis

algorithm, and the severe sepsis order set was implemented if necessary

based on the patient's criteria and/or physician's discretion. Data was

collected for a 3-month period after the beginning of the study. An

attempt was made to obtain a similar sample size as the pre-

implementation period. The severe sepsis screening tool, severe sepsis

algorithm, and severe sepsis order sets were approved by the Medical

Executive Committee to be utilized as a standard of care on all patients in

the designated medical surgical unit. Post-Intervention: Following the

completion of the data collection period, de-identified data was extracted

from Data Diver using patients who screened positive for systemic

inflammatory response syndrome (SIRS), sepsis, severe sepsis (including

42

urosepsis), an/or septic shock and that the severe sepsis algorithm was

utilized. Once again specific data points were collected including

transfers to higher level of care and hospital length of stay. A comparison

was appraised with results obtained from the pre-implementation data.

4. Evaluate the sensitivity and specificity of a severe sepsis tool. Evaluation of the

Severe Sepsis Tool: In order to evaluate the sensitivity and specificity of the

Severe Sepsis Tool utilized at the organization, information technology (IT)

generated a report of collected data within the last year regarding medical-surgical

patients in the unit where the pilot was implemented: total number of patients

with discharge diagnosis of sepsis criteria (systemic inflammatory response

syndrome, sepsis, severe sepsis (including urosepsis), an/or septic shock) based

on ICD-9 codes was collected. All patients who screened positive more than one

time for sepsis criteria during hospitalization and all patients who screened

negative during hospital stay were accounted. A receiver operating curve (ROC)

and the respective area under the curve were calculated. Utilizing a 2x2 design,

the sensitivity and specificity of the tool were obtained.

43

Budget

Project Budget Total Budget $2065.00

Nurses Research Department

Education/Training (40 nurses at $26/hr)

$1040.00 Education component not part of research

Power Analysis $80.00 Accounted during regular

work hours

ARNP Education (5 ARNPs at $45/hr)

$225.00 Education component not part of research

Data Analysis pre-

implementation

$160.00

Total $1265.00 Accounted during regular

work hours

Data Analysis post-

implementation

$160.00

Information Technology Department Materials

IT Sepsis Report ($15/hr) pre-

implementation

$150.00 Paper/Copies $200.00

IT Sepsis Report ($15/hr) post-

implementation

$150.00

Total $300.00

Budget Analysis

Nurses

Education/Training $1040.00

ARNP Education $225.00

A total of 40 registered nurses were trained on the use of the protocol. The average

salary of these nurses was $26/hour. The in-service was for 30-45 minutes at a rate of

$26 per nurse. A total of five ARNPs were trained on the use of the protocol. The

44

average salary of these ARNPs was $45/hour. The in-service was for 30-45 minutes at a

rate of $45 per ARNP.

Information Technology

Information technology provided pre- and post-implementation data sets, an

average of $15/hour was allocated, combining a total of $ 300 for their collaboration in

the project.

Research Department

The research department provided a pre- and post-data analysis, allocating a total

of $300 for their collaboration in the project.

Timeline of Project

The timeline of events designated to the capstone project were as follows.

Dates: Month/Year Description of Event

8/14-1/15 IRB approval from both entities: Nova and the organization and

complete all the necessary paperwork to implement project,

including approval from different organizations’ committees:

Medical Executive Committee, Nursing Board

1/15 Education component and pre-implementation data collection: All

staff will be educated on the proposed project, and retrospective

data will be obtained with attention to selected data points:

Transfer to higher level of care and length of stay.

2/15-5/15 Implementation of algorithm and sepsis tool and data collection in

45

pilot unit. All data will be collected retrospective after the

conclusion of the data collection period.

5/15 Analysis of collected data: Independent t-test and chi square.

Analytics will be utilized to evaluate the collected data by the

organization’s research department.

6/15 Evaluation of results

Determination of Sample Size: Power Analysis

A statistical power analysis was conducted to calculate the appropriate sample

size given the level of significance (α error probability), power (1-β error probability),

and effect size. There are at least three ways to approximate the necessary values to

conduct a power analysis when performing t-tests: literature review, pilot study, and

Cohen’s recommendations. Although literature reviews sometimes can be powerful,

many times, the required effect sizes to properly estimate the power analysis are not

provided in the articles; therefore making literature reviews a difficult choice to formulate

a power analysis. Even though this study is a pilot, and pilot studies are known to

provide rough estimates of the effect size and variability of the measurements, it was

determined that the best approach was to utilize Cohen’s recommendation. Pilot studies

usually are required to estimate small sample size, making it difficult to evaluate

significance (IDRE, 2014). In order to control Type I error, α value was set at 0.05 and to

control Type II error β value was set at 0.95. Anticipated effect size (Cohen’s d) was set

as medium effect, 6% of the variance: d = 0.5 (Cohen, 1988). Based on the above values,

an anticipated sample size of 128 participants was necessary for each group (N = 256). In

46

order to calculate power analysis for chi-square, it was determined that the most efficient

approach was to utilize a power table chi-square as a function of population values of

Cramer’s V. Anticipated α value was set at 0.05, with power set at 0.80 and V statistics

set at 0.30. Based on the above values, an anticipated sample size of 87 participants was

necessary for each group (N = 174).

Statistical Tests

In order to analyze the proposed research questions of determining the sensitivity

and specificity of the severe sepsis tool and evaluating if the implementation of a severe

sepsis algorithm in a medical surgical ward will decrease patient transfer to higher level

of care and hospital length of stay, multiple statistical tests were conducted. First,

objective #4 sought to evaluate the sensitivity and specificity of the tool by utilizing a

receiver operating curve (ROC) and the respective area under the curve. Objective #3

sought to compare the mean differences between septic patients transfers to higher level

of care and hospital length of stay after an implementation; therefore, a chi-square test

and two-tailed independent t-test were utilized.

Outcome Measures

The outcomes of this project were evaluated using the following measures:

Outcome 1: Generate administration and management support for the project.

This objective was measured by the organization’s approval on the utilization of the

severe sepsis bundle during the study. This was implemented as a pilot study to allow for

questions, identification, and correction of any issue that may arise on the part of the

implementation.

47

Outcome 2: Evaluate septic patients hospital length of stay and patients transfers to

higher level of care on medical-surgical unit prior to implementation. Following IRB

approval, a request was made to information technology (IT) for two reports. The first

report contained de-identified data for all patients admitted to the designated medical

surgical unit over a 3-month period and who presented with a diagnosis of systemic

inflammatory response syndrome (SIRS), sepsis, severe sepsis (including urosepsis),

and/or septic shock based on ICD-9 codes. The second report contained de-identified

data for all patients admitted to the designated medical surgical unit over a 3-month

period and who presented with a diagnosis other than systemic inflammatory response

syndrome (SIRS), sepsis, severe sepsis (including urosepsis), and/or septic shock based

on ICD-9 codes. De-identified data obtained from those reports specifically pertaining to

transfers to higher level of care and length of stay were calculated.

Outcome 3: Develop and implement a severe sepsis bundle (tool, algorithm, and order

set) for medical surgical units. After the development of a severe sepsis bundle

accounting for a tool, algorithm, and order sets, proper education was provided to the

staff. Although the bundle was approved by the organization, the nurses played an

important role in properly utilizing the bundle. It required excellent critical thinking

skills for proper identification, implementation, and treatment of septic patients from

bedside nurses. This objective was measured by comparing the pre-implementation and

post-implementation data evaluating septic patients hospital length of stay and septic

patients transfers to higher level of care.

Objective 4: Evaluating the sensitivity and specificity of the severe sepsis tool utilized

at the organization. In order to evaluate this objective, the total number of admissions

48

from pilot unit during a specific time frame was collected. The number of those admitted

patients with a discharge diagnosis of sepsis was attained. Out of those patients with a

discharge diagnosis of sepsis, all patients who had at least one positive screening for

severe sepsis was counted. Lastly, of those patients with a discharge diagnosis of sepsis

who screened negative for sepsis was also counted. The results yielded the sensitivity and

specificity of the tool.

Validity and Reliability of the Sepsis Screening Tool

Following the established SSC national guidelines, a screening tool was

developed by a group of critical care leaders including nurse executives, managers, and

educators. The rationale of the screening tool was to facilitate bedside nurses with ways

to identify and treat septic patients before becoming hemodynamically unstable. The

advantage of early identification is that evidence-based treatment options can be initiated

promptly and consistently, thus improving patients’ outcomes (Cabrera, 2010). The

screening tool in practice was validated by five critical care professional educators, each

with more than 50 years of experience in the nursing field. Each individual step in the

tool, as well as the overall tool, was assessed for readability, relevance, and

appropriateness. The tool was also validated for accuracy and congruency with

established national guidelines (Cabrera, 2010). The tool has also been utilized at the

community hospital ICU department as well as in multiple research projects within the

organization. A reliability study was conducted with a convenience sample of 40 critical

care nurses with multiple levels of experience to practice different sepsis scenarios

utilizing the tool. The study revealed consistency in the nurses’ categorization of the

screening tool, demonstrating the reliability and validity of the tool. The tool has been

49

also adopted into other ICUs within the organization obtaining the same reliability scores

(Cabrera, 2010).

Summary

This study was a quasi-experimental pre- and posttest research that utilized

multiple phases in the implementation of an evidence-based sepsis tool and sepsis

algorithm. The study evaluated a specific medical- surgical sepsis algorithm that can

potentially reduce transfer to higher level of care and hospital length of stay. A medical-

surgical unit was selected for the study because of its high sepsis length of stay in

comparison to other units in the organization. The patients’ rights and confidentiality

were protected by obtaining IRB approval from both the organization and Nova

Southeastern University. The study’s outcomes were measured by evaluating the

achievement of each objective.

50

Chapter 4: Results and Discussion

The purpose of this pilot study was to employ a severe sepsis tool (Appendix C)

already utilized at a community hospital in Miami, Florida, newly implement a severe

sepsis algorithm (Appendix D) for early identification of severe septic patients in a

medical-surgical non-ICU and severe sepsis order set (Appendix E), thus potentially

reducing patient transfers to higher levels of care and hospital length of stays. This

chapter will provide an overview of the results obtained after the implementation of the

capstone project, specifically focusing on each objective and data points. A discussion

will be provided on the implications for nursing practice, healthcare delivery, and

healthcare policy as well as areas for future research.

This project employed a non-randomized quasi-experimental design, in which a

non-equivalent group comparison was made between homogenous sample utilizing the

International Classification of Diseases (ICD-9) Coding for SIRS, sepsis, severe sepsis,

and septic shock (Appendix G). A pre test-posttest design was utilized to evaluate the

implementation of a severe sepsis bundle specifically designed for medical surgical units.

Hospital length of stay and transfers to higher level of care were the measured outcomes.

The sensitivity and specificity of the utilized sepsis tool in the pilot study were also

calculated. Patients were treated in a 50-bed medical-surgical ward at a 680-bed

community hospital in Southeast Florida.

51

Descriptive Results

Objective 1

Administration and management support for the project was obtained by

acquiring a letter of support from the organization’s administration and management

department. Consent to implement the project on the selected medical-surgical unit was

also validated with a letter from the nurse manager (Appendix B).

Objectives 2 and 3

For the evaluation of septic length of stay, a convenience sample of patients

admitted to the medical-surgical unit before and after implementation of the severe sepsis

bundle (screening tool, algorithm, and order set) was obtained utilizing ICD-9 coding and

discharge date during a period of 6 months. During the project's implementation phase, a

total of 264 participants were identified with sepsis ICD-9 code. One hundred and thirty

one patients (n = 131) were identified as having met criteria for sepsis syndrome over a 3-

month period prior to implementation. The post-implementation group included 121

patients (n = 121) who met criteria for sepsis syndrome over a period of 3-month. A total

of 12 patients were removed from the database due to missing data (seven from pre-

implementation, five from post-implementation group). Table 1 provides a breakdown of

the number of patients in each group.

52

Table 1 Description of Comparison Groups

Group Statistics Pre/Post N Mean Std. Deviation Std. Error Mean

LOS 1 131 8.53 11.138 .973 2 121 7.26 7.532 .685

For the evaluation of sepsis transfers to higher level of care from the selected

medical-surgical ward, a convenience sample was not able to be obtained secondary to

insufficient numbers of transfers. The intended sample size (n = 174) based on the power

analysis for Cramer’s V was set at 0.05, with power set at 0.80 and V statistics set at 0.30

was not accomplished due to small number of transfers from both the pre and post

groups.

Objective 4

Once data had been entered and compiled for the evaluation of the sensitivity and

specificity of a severe sepsis screening tool, IT generated a report for the time frame of

April 2014 – April 2015. The report contained: total number of patients admitted to the

medical surgical ward (N = 1,555): of those patients, total number with discharge

diagnosis of sepsis (n = 193), total number who screened positive greater than one time

during hospital stay (n = 78), and total number who screened negative during hospital

stay (n = 110); there were five missing cases. Receiver operating curve (ROC) and the

respective area under the curve were calculated. Utilizing a 2x2 design, the sensitivity

and specificity of the tool was calculated.

53

Statistical Results

Objective 1. Once administration and management support for the project was

obtained, there was no need to perform any statistical analysis for objective one.

Objectives 2 and 3. The purpose of these two objectives were to develop and

implement a severe sepsis bundle specifically designed for medical-surgical wards with

the intention of decreasing sepsis length of stay and sepsis transfer to higher level of care.

A convenience sample of patients admitted to the medical-surgical unit before and after

implementation of the severe sepsis bundle (screening tool, algorithm, and order set) was

obtained. Once data had been entered and compiled, a normal distribution was found

among the two groups in the length of stay calculation. Subsequently, independent t-

tests, and Levene's Test for Equality of Variances were used to analyze the data. The

tests revealed that there was no statistical significance (p = .291) in terms of length of

stay between the pre- and post-implementation groups. Table 2 provides independent t-

test results for each group.

54

Table 2 Independent t-Test Results for Each Group

Independent Samples Test Levene’s Test for Equality of Variances

F Sig LOS Equal variances assumed Equal variances not assumed

4.255 .040

Independent Samples Test t-Test for Equality of Means t df Sig. (2-

tailed) Mean

Difference LOS Equal variances assumed Equal variances not assumed

1.058 1.074

250 229.607

.291 .284

1.278 1.278

Independent Samples Test t-Test for Equality of Means Std. Error

Difference Lower Upper

LOS Equal variances assumed Equal variances not assumed

1.208 1.190

-1.100 -1.066

3.657 3.623

Despite the lack of statistical significance in the findings, it is important to recognize that

there was, indeed, clinical significance. The study revealed that in fact there was a

decrease in sepsis length of stay, from pre implementation of 8.53 days to post

implementation of 7.26, however the data was not statistically significance.

55

For the evaluation of sepsis transfer to higher level of care, the data collected

revealed that during the 6-month period of the study, there were zero transfers of septic

patients. Tabulating the information obtained and using the International Classification

of Diseases (ICD-9) Coding for SIRS, sepsis, severe sepsis and septic shock, the data

discovered demonstrated that during the study implementation phase there were only five

patients (pre n = 4, post n = 1) transfers to either the intensive care unit or a progressive

care unit from the medical surgical ward that the study was implemented; however, the

transferred patients did not have a sepsis diagnosis. Therefore, it was not possible to

calculate chi square test to determine whether there was a significant association between

septic transfers to higher level of care and the implementation of the sepsis bundle.

Objective 4. A total of 2,840 patients were screened utilizing the severe sepsis

screening tool. The receiver operating curve (Figure 2) and the respective area under the

curve were calculated. Utilizing a 2x2 design, the sensitivity and specificity of the tool

were calculated. Four hundred and fifty three (n = 453) patients were diagnosed with

sepsis criteria at time of discharge. Two hundred and fifty three (n = 253) screened

positive for sepsis syndrome. The study yielded a sensitivity of 55.8%. Two thousand

three hundred and seventy six (n = 2376) patients were not diagnosed with sepsis

syndrome at the time of discharge and screened negative for sepsis. The study yielded a

specificity of 99.54%. The positive predictive value (PPV) of the tool was estimated at

95.83%, negative predictive value (NPV) was estimated at 92.83%, and disease

prevalence was 15.95% (Table 5). Area under the receiver operating curve was 0.777,

meaning according to area under the ROC curve (AUROCC), the results will be

classified as a fair test. However, it is important to recognize that the positive predictive

56

value (PPV) of the tool was 95.83%, meaning that the tool has a high probability of being

correct when the sepsis disease is actually present.

Table 3 2x2 Tabulation Design

Sepsis

Present

Sepsis Absent

Positive 253

True Positive

11

False Positive

Negative 200

False Negative

2376

True Negative

Table 4 Area Under the Curve

Area 95% Wald Confidence Limits 0.7769 0.7540 0.7999

57

Table 5 Sensitivity and Specificity

Results Value 95% CI Sensitivity 55.85% 51.14% to 60.48% Specificity 99.54% 99.18% to 99.77% Positive Likelihood Ratio

121.19 66.83 to 219.79

Negative Likelihood Ratio

0.44 0.40 to 0.49

Disease Prevalence 15.95% (*) 14.62% to 17.35% Positive Predictive Value

95.83% (*) 92.67% to 97.90%

Negative Predictive Value

92.24% (*) 91.13% to 93.24%

Strengths and Limitations of the Project

Strengths

This capstone project was designed to not only validate the severe sepsis tool

sensitivity and specificity but also develop and implement a severe sepsis bundle to

potentially reduce sepsis length of stay and sepsis transfers to higher level of care from a

medical surgical ward. This capstone project demonstrated multiple strengths that were

significant to the data collection methodology and obtained results. One important

strength of the study was that despite the lack of statistical significance in the findings, it

is important to recognize that there was, indeed, clinical significance. Nursing staff as

well as physicians who participated were educated and were more proactive in the

identification and management of septic patients. Lack of knowledge and awareness of

practice changes have been identified as obstacles to implementation of evidence-based

practice (Rubenfeld, 2004). Even though nurses’ perceptions of the project were not

obtained, the education process of this project was, in fact, a strength. Another strength

was that although not significant, there was indeed a reduction in the length of stay from

58

the pre implementation group to the post implementation group. It is well documented in

the literature that length of stay is directly correlated with hospital operating cost;

therefore, by reducing the length of stay of septic patients, it can assume that there was

also a reduction in operational expenses.

Another important strength of the study was that there was zero transfer of septic

patients to an intensive care unit or progressive care units, which practitioners can

speculate that could be because of the excellent performance that clinicians are doing on

a daily basis in identifying and treating septic patients. Although the power analysis for

this objective was calculated, given the fact that there was zero transfers, made it

impossible to tabulate the chi square score. Following the same ideology, although not a

data point in this study, hospital operational cost were probably reduced by eliminating

septic transfers to higher level of care.

Regarding the severe sepsis screening tool, it was determined that the tool was

excellent in identifying patients who were not septic, but it was weak in identifying

patients who were septic. This finding is valued as a strength because it provides a

foundation for ways to improve and calibrate the tool. A single phase retrospective study

demonstrated that the severe sepsis screening tool utilized at a community-based hospital

in Miami, Florida has a sensitivity value of 55.85% and a specificity value of 99.54%

when evaluating medical surgical patients. These results indicate the tool is accurate in

detecting patients who are not septic; however, it was not reliable in identifying patients

who are truly septic. Nevertheless, the study concluded a positive predictive value of

95.83%.

59

Limitations

There were limitations to the clinical practice capstone project. The study utilized

convenience sample instead of randomization. The study also focused on only two

essential data points: sepsis length of stay and sepsis transfers to higher level of care.

Although simple, the data collected on this capstone study was not utilized to evaluate

any other data points, such as demographics, mortality, or APACHE scores. The study

also did not focus on any qualitative perspective; therefore, no data was collected on the

perceptions of nurses or physicians evaluating their understanding of the sepsis bundle or

screening tool.

Another important limitation was the sample size. Even though a power analysis

was conducted, the minimum amount of patients required to be statistically significant

was 256; however, only 252 patients were tabulated because 12 patients’ data were

missing. Additional limitations to this study were the time frame and the unit utilized.

The developed sepsis bundle was implemented in only one medical surgical ward for a

period of 3 months. Data obtained was then compared to 3 months information from pre

implementation time. However, during the implementation of this project, there were

serious time constrictions, mainly because of IRB and university approval. It is believed

that 6 months might have been too short of a time frame to adequately analyze the impact

of the sepsis bundle. Perhaps extending the time frame of the project's implementation

phase could have demonstrated a statistically significance on sepsis length of stay and

transfers to higher level of care. The utilization of only one medical surgical unit, though

convenient, could have impacted the project's results. It is well established that not all

medical surgical wards are the same, especially when describing the level of acuity and

60

complications. Nevertheless, the findings from this project demonstrate that early

identification and early interventions of sepsis bundles are effective in delivery of

evidence-based care.

When evaluating the sensitivity and specificity of the sepsis tool, multiple

limitations were identified. Some of the identified limitations to this study included:

sample size, nurse’s implementation and understanding of the utilization of the tool, lack

of education regarding the tool before implementation and/or inability to accurately

diagnose patients with sepsis syndrome on admission. Even thought 2,840 patients were

included in this portion of the study, the results obtained could have been improved with

a larger number of participants. Perhaps obtaining similar sample sizes from different

medical surgical wards throughout the organization could potentially accentuate the

reliability and validity of the tool as well as improve the sensitivity. Also, a lack of

compliance with the tool could have affected the results, mainly due to lack of knowledge

on the proper utilization of the tool, given that the education perspective only occurred

during the last 3 months of the study. Despite the results obtained, the study did provide

areas to improve.

Implications of the Study

Implications for Nursing Practice

Nurses play a key role in identifying patients with severe sepsis or septic shock

and advocating for these patients. The findings from this project demonstrated that

improving sepsis awareness in medical surgical units by simultaneously educating the

staff and providing better tools to both identify and treat septic patients are essential in

the patients’ prognosis and recovery. In spite of the growing medical and economic

61

burden of sepsis, through early identification and treatment of septic patients, nurses can

ensure adequate prevention and therapy for all septic patients. Early sepsis identification

has demonstrated improvement in decreasing severe sepsis incidences on medical

surgical units and decreasing transfer to higher level of care and lengthy hospital stays.

After the implementation of this capstone project, nurses in 4 Tower were able to

identified and rapidly treat septic patients, while collaborating with the physician in a

more proactive manner. The creation of this severe sepsis algorithm and sepsis bundle

specifically designed for medical surgical units allowed the nursing staff to provide

adequate early goal directed therapy. Even though some of the results obtained were not

statistically significant, the sepsis bundle assisted physicians and practitioners to better

diagnose patients with sepsis. Additionally, this pilot study added research to the body of

knowledge allocated with sepsis.

Implications for Healthcare Delivery

The intent of this capstone project was to positively impact the healthcare delivery

system by improving awareness on the topic of sepsis in medical surgical units. Through

increased knowledge of sepsis, organizations and practitioners are able to confidently

influence the healthcare delivery system on the subject of sepsis. One of the most

important recommendations of the Surviving Sepsis Campaign (SSC) of 2012 is the

utilization of sepsis tools and bundles for early sepsis identification and early goal-

oriented therapy, advocating healthcare institutions and clinicians to be actively engaged

in the sepsis movement. The results of this project demonstrated that early identification

and early interventions of sepsis bundles are effective in delivery of evidence-based care.

62

The severe sepsis bundle should be incorporated into all medical surgical wards within

the organization and even be adapted to other sister institutions.

The Center for Medicare & Medicaid Services (CMS) and Quality Net has

recently implemented the Hospital Inpatient Quality Reporting Program Measures

International Classification of Diseases, 10th Edition, adding sepsis as a core measure

(Quality Net, 2015). This concept of implementing a sepsis bundle will no longer be a

privilege but a mandate; therefore, the sepsis bundle utilized in this project can be utilized

to fulfill the requirements of CMS.

Implications for Healthcare Policy

Sepsis, much like other healthcare issues, has become more conspicuous among

policymakers. Due to the increased prevalence and high mortality rate, now more than

ever is the perfect time to implement mandates and guidelines to improve the nation’s

sepsis outcomes. This capstone project will add to the existing body of knowledge on the

topic of sepsis at a local and regional level. As stated previously, in 2013, the state of

New York became the first state in the nation to mandate for every hospital to adopt an

evidence-based protocol for the early diagnosis and management of sepsis (Cuomo,

2013). Following the SSC recommendations and the endorsement of the National

Quality Forum, New York hospitals are now required to utilize screening tools for early

recognition of patients with sepsis, severe sepsis, and septic shock, in addition to early

goal-oriented therapy for the management of septic patients. Through the implementation

of this project and others, Florida could become a proactive sepsis state such like New

York. This project's design and methodology can be adapted to implement other research

projects to continue to improve the nation’s healthcare system.

63

Future Research

The preponderance of the available literature on the topic of sepsis surrounds the

management of severe septic and septic shock patients in ICUs. Severe sepsis has been

largely studied in the intensive care unit, and limited research is available regarding

sepsis identification and treatment in medical-surgical units. Increasing the body of

knowledge and information through evidence-based research and study findings should

be accentuated as essential objectives.

The presumption of obtaining information from clinical practice research is to

subsequent provide answers based on findings/conclusions of gap analysis; therefore, this

project will add to the existing body of knowledge, exclusively implementing important

strategies in medical surgical units. Areas for future research regarding sepsis could be to

perform correlational studies to identify if a relationship exists between various variables,

specifically exploring the limitations of this study. Perhaps exploring larger, more

assorted populations could potentially provide more favorable results. Replication

studies could also be another area of exploration for future research, given that no two

sites are alike. Conceivably, the same study implemented at different locations can

establish generalizability of results.

As healthcare continues to change and new guidelines are implemented regarding

sepsis, organizations will be forced to embrace the power of research to achieve desired

outcomes. Early recognition of sepsis and improved therapies to manage the multi-organ

dysfunction that frequently follows sepsis pathophysiology remain majorly unmet

medical needs. The interaction between virulence factors of the pathogen and the host

defense system are decisive implications in sepsis patients, especially when improving

64

outcomes (Singer, 2013). Therefore, future research should be inclined towards

connecting pathogen identification with prompt therapeutic intervention, steps in which

nurses play a crucial role. This capstone study could provide the foundation for future

research in sepsis in medical surgical areas to improve the clinical practice and health

paradigm.

Summary

Severe sepsis continuous to strike about 750,000 Americans annually, with a

mortality rate estimated at 28% to 50%. It is approximated that sepsis deaths in the U.S.

accounts for more than the deaths from prostate cancer, breast cancer, and AIDS

combined (Dellinger, et al., 2013). Sepsis is on the rise due to aging population,

increased longevity of people with chronic diseases, spread of antibiotic-resistant

organisms, and increase in invasive procedures. Standardized guidelines have been

developed with the intention of providing clarity in the treatment and management of

sepsis.

Severe sepsis has been largely studied in the intensive care unit. Limited research

is available regarding sepsis identification and treatment in medical-surgical units. This

capstone project focused on the implementation and evaluation of a sepsis bundle

specifically designed for medical surgical units. Since the initial development of the

Surviving Sepsis Campaign guidelines published in 2004 outlining the importance of

early identification and management of severe sepsis, there has been an absolute

disregard of the management of septic patients in medical surgical units, formulating a

problem among healthcare providers including nurses on the proper steps to diagnose and

treat septic patients.

65

This pilot study employed a severe sepsis bundle in a medical-surgical non-ICU,

specifically looking at patient transfer to higher levels of care and hospital length of

stays. The study demonstrated no statistical significance between the sepsis bundle

implementation and reduction in sepsis transfers or sepsis hospital length of stay. The

study’s results could have been obtained due to multiple limitations to the study. These

limitations could be potential opportunities for future research. Nevertheless, the study

demonstrated that early identification and early interventions of sepsis bundles are

effective in delivery of evidence-based care. The severe sepsis bundle should be

incorporated into all medical surgical wards within the organization and even be adapted

to other sister institutions.

66

References

Anderson, R., & Schmidt, R. (2010). Clinical biomarkers in sepsis. Front Bioscience (Elite Edition), 2(5), 504-520.

Angus, D. C., Linde-Zwirble, W. T., Lidicker, J., Clermont, J., Carcillo, J., & Pinsky, M. R. (2001). Epidemiology of severe sepsis in the United States: Analysis of incidence, outcome, and associated costs of care. Critical Care Medicine, 29(7), 1303-1310.

Baptist Hospital of Miami. (2014). Outcomes measures. Quality, Services and Safety. Retrieved from https://baptisthealth.net/en/facilities/baptist-hospital-

miami/pages/about.aspx Burney, M., Underwood, J., McEvoy, S., Nelson, G., Dzierba, A., Kauari, V., & Chong, D. (2012). Early detection and treatment of severe sepsis in the emergency department: Identifying barriers to implementation of a protocol-based approach. Journal of Emergency Medicine 38(6), 512-517. Cabrera, A.(2010). Efficacy of early identification and intervention bundles in ICU

patients identified with severe sepsis and septic shock (Capstone project). Barry University, Miami Shores, FL.

Carter, C. (2007). Implementing the severe sepsis care bundles outside the ICU by outreach. Nursing Critical Care, 12(5), 225-230. Centers for Medicare & Medicaid Services. (2014). Quality measures. Retrieved from

http://www.cms.gov/Medicare/Quality-Initiatives-Patient-Assessment- Instruments/QualityMeasures/index.html

Clearly, P. D., & O'Kane, M. E., (2013). Evaluating the quality of health care. E-source Behavioral and Social Sciences Research. Retrieved from http://www.esourceresearch.org/Portals/0/Uploads/Documents/Public/Cleary_Full Chapte r.pdf

Cuomo, A. M. (2013). Governor Cuomo announces New York state to lead the nation in fighting sepsis – the #1 killer in hospitals – and make major improvements in pediatric care through "Rory's Regulations." Retrieved form http://www.governor.ny.gov/press/012913-nys-lead-nation-fighting-sepsis

Dellinger, R. P., Levy, M. M., Rhodes, A., Annane, D., Gerlach, H., Opal, S. M, … Moreno, R. (2013). Surviving sepsis campaign: International guideline for management of severe sepsis and septic shock 2012. Critical Care Medicine, 41(2), 580-637.

Donabedian, A. (1966). Evaluating the quality of medical care. The Milbank Memorial Fund, 44(3), 166-206.

67

Donabedian, A. (1980). Explorations in quality assessment and monitoring vol. 1. The definition of quality and approaches to its assessment. Ann Arbor, MI: Health Administration Press. Donabedian, A. (1988). The quality of care. How can it be assessed? The Journal of the American Medical Association, 260(12), 1743-1748. Donabedian, A. (1990). The seven pillars of quality. Archives of Pathology &

Laboratory Medicine, 114(11), 1115-1118. Donabedian, A. (1998, April 16). Interview by Edward Berkowitz. History of Health

Services Research Project Interview with Avedis Donabedian. National Information Center on Health Services Research and Health Care Technology (NICHSR). Ann Arbor, Michigan:Health Press

Donabedian, A. (2005). Evaluating the quality of medical care. The Milbank Quarterly, 83(4), 691-729. Dougherty, M. L. (2012). Central line-associated bloodstream Infection prevention in the

long-term acute care setting (Doctoral dissertation). Grand Valley State University, Allendale, MI. Retrieved from http://scholarworks.gvsu.edu/cgi/viewcontent.cgi?article=1004&context=dissertat ions

Durthaler, J. M., Ernst, F. R., & Johnston, J. A. (2009). Managing severe sepsis: A

national survey of current practices. American Journal of Health-System Pharmacology 66, 45-53.

El Haj, H. I., Lamrini, M., & Rais, N. (2013). Quality of care between Donabedian model and ISO9001V2008. International Journal for Quality Research 7(1), 17-30.

Elixhauser, A., Friedman, B., & Stranges, E. (2011). Septicemia in U.S. hospitals, 2009. Healthcare Cost and Utilization Project: Statistical Brief 122. Retrieved from http://www.hcup-us.ahrq.gov/reports/statbriefs/sb122.pdf

Gamez-Diaz, L. Y., Enriquez, L. E., Matute, J. D., Velasquez, S., Gomez, I. D., Toro, F., … Jaimes, F. (2011). Diagnostic accuracy of HMGB-1, s-TREM-1, and CD64 as markers of sepsis in patients recently admitted to the emergency department. The Society for Academic Emergency Medicine, 18(8), 807-815.

Gibot, S., Bene, M.C., Noel, R., Massin, F., Guy, J., Cravoisy, A., … Charles, P.(2012).

Combination biomarkers to diagnose sepsis in the critically ill patients. American Journal of Respiratory and Critical Care Medicine, 186, 65-71.

Giuliano, K. K., Lecardo, M., & Staul, L. (2011). Impact of protocol watch on

compliance with the surviving sepsis campaign. American Journal of Critical Care 20(4), 313-321.

68

Hall, M. J., Williams, S. J., DeFrances, C, J., & Golosinskiy, A. (2011). Inpatient care for septicemia or sepsis: A challenge for patients and hospitals. Centers for Disease Control and Prevention National Center for Health Statistics, 62, Retrieved from http://www.cdc.gov/nchs/data/databriefs/db62.pdf Hoeboer, S. H., Alberts, E., Van den Hul, I., Tacx, A. N., Debets-Ossenopp, Y. J., &

Groeneveld, J. (2012). Old and new biomarkers for predicting high and low risk microbial infection in critically ill patients with new onset fever: A case for procalcitonin. Journal of Infection, 64, 484-493.

Hohn A., Schroeder, S., Gehrt, A., Bernhardt, K., Bein, B., Wegscheider, K., & Hochreiter, M. (2013). Procalcitonin-guided algorithm to reduce length of antibiotic therapy in patients with severe sepsis and septic shock. BioMedical Central Infectious Diseases, 13(158), 1-9.

Institute of Medicine. (2001). Crossing the quality chasm: A new health system for the 21st century. Washington D.C.: National Academy Press. Jekarl, D. W., Lee, S. Y., Lee, J., Park, Y. J., Kim, Y., Park, J. H., … Choi, S.P. (2013).

Procalcitonin as a diagnostic marker and IL-6 as a prognostic marker for sepsis. The Lancet Infectious Disease, 75(4), 342-347.

Kopterides, P., Siempos, I. I., Tsangaris, I., Tsantes, A., & Armaganidis, A. (2010).

Procalcitonin-guided algorithms of antibiotic therapy in the intensive care unit: A systematic review and meta-analysis of randomized controlled trials. Critical Care Medicine, 38(11), 2229-2241.

Lawson, E. F., & Yazdany, J.(2012). Healthcare quality in systemic lupus erythematosus: using Donabedian’s conceptual framework to understand what we know. International Journal of Clinical Rheumatology, 7(1), 95-107.

Levy, M. M., Dellinger, R. P., Townsend, S. R., Linde-Zwirble, W. T. , Marshall, J. C.,

Bion, J., … Angus, D.C. (2010). The surviving sepsis campaign: Results of an international guideline-based performance improvement program targeting severe sepsis. Critical Care Medicine, 38(2), 367-374.

Lipinska-Gediga, M., Mierzchala, M., & Durek, G. (2012). Pro-atrial natriuretic peptide (pro-ANP) level in patients with severe sepsis and septic shock: Prognostic and diagnostic significance. Clinical and Epidemiological Study, 40, 303-309. Marwick, C., & Davey, P. (2009). Care bundles: the holy grail of infectious risk

management in hospital? Current Opinion in Infectious Diseases, 22, 364-369. McClelland, H., & Moxon, A. (2014). Early identification and treatment of sepsis. Nursing Times, 110(4), 14-17.

69

Nelson, D. P., LeMaster, T. H., Plost, G. N, & Zahner, M. L. (2009). Recognizing sepsis in the adult patient. American Journal of Nursing, 109(3), 40-45. Oh, J. S., Kim, S. U., Oh, Y., Choe, S. M., Choe, G. H., Choe, S. P., … Park, K. N.

(2009). The usefulness of the semiquantitative procalcitonin test kit as a guideline for starting antibiotic administration. American Journal of Emergency Medicine, 27, 859-863.

Pierce, J. D., McCabe, S., White, N., & Clancy, R. L. (2012). Biomarkers: An important clinical assessment tool. American Journal of Nursing, 112(9), 52-58. Pierrakos, C., & Vincent, J. (2010). Sepsis biomarkers: A review. Critical Care, 14(15). Retrieved from http://ccforum.com/content/14/1/R15 Polit, D. F., & Beck, C. T. (2012). Nursing research: Generating and assessing evidence for nursing practice 9th edition. Philadelphia, PA: Wolters Kluwer. QualityNet (2014). Specifications Manual, Version 4.4a. Retrieved from

https://www.qualitynet.org/dcs/ContentServer?c=Page&pagename=QnetPublic% 2FPage%2FQnetTier4&cid=1228773989482.

Rhode, J. M., Odden, A. J., Bonham, C., Kuhn, L., Malani, P. N., Chen, L. M., …

Iwashyna, T. J. (2013). The epidemiology of acute organ system dysfunction from severe sepsis outside of the intensive care unit. Journal of Hospital Medicine, 8(5), 243-247. doi: 10.1002/jhm.2012.

Rivers, E. P., & Ahrens, T. (2008). Improving outcomes for sever sepsis and septic shock: Tools for early identification of at-risk patients and treatment protocol implementation. Critical Care Clinics, 23, S1-S47. Rubenfeld, G. (2004). Translating clinical research into clinical practice in the intensive care unit: the central role of respiratory care. Respiratory Care, 49, 837-843. Sankar, V., & Webster, N. R. (2013). Clinical application of sepsis biomarkers. Journal

of Anesthesia, 27, 269-283. doi: 10.1007/s00540-012-1502-7 Shaaban, H., Daniel, S., Sison, R., Slim, J., & Perez, G. (2010). Eosinopenia: Is it a good marker of sepsis in comparison to procalcitonin and c-reactive protein levels for patients admitted to a critical care unit in an urban hospital? Journal of Critical Care, 25, 570-575. Singer, M. (2013). Biomarkers in sepsis. Current Opinion in Pulmonary Medicine, 19(00), 1-5. doi: 10.1097/mcp.0b13e32835f1b49.

70

Smith, E. L., Rice, K. L., & Winterbottom, F. (2012). Nurses' critical role in identifying sepsis and implementing early goal-directed therapy. The Journal of Continuing Education in Nursing, 43(6), 247-248. Tang, B., Eslick, G., Craig, J., & McLean, A. (2007). Accuracy of procalcitonin for

sepsis diagnosis in critically ill patients: Systematic review and meta-analysis. The Lancet Infectious Disease, 7(3), 210-217.

Tazbir, J. (2012). Early recognition and treatment of sepsis in the medical-surgical setting. Medical Surgical Nursing, 21(4), 205-208. Thiel, S.W., Asghar, M. F., Micek, S. T., Reichley, R. M., Doherty, J. A., & Kollef, M.H. (2009). Hospital-wide impact of a standardized order set for the management of bacteremic severe sepsis. Critical Care Medicine, 37(3), 819-24. doi: 10.1097/CCM.0b013e318196206b Tromp, M., Tijan, D. H. T., van Zanten, A. R. H., Gielen-Wiffels, S. E. M., Goekoop, G.

J. D., van den Boogaad, M., … Pickkers, P. (2011). The effects of implementation of the surviving sepsis campaign in the Netherlands. Netherlands Journal of Medicine, 69(6), 292-298.

Wacker, C., Prkno, A., Brunkharst, F. M., & Schlattmann, P. (2013). Procalcitonin as a

diagnostic marker for sepsis: a systematic review and meta-analysis. The Lancet Infectious Disease, 13, 426-435.

Wang, H., E., Shapiro N., I., Angus, D. C., & Yealy, D.,M. (2007). National estimates of severe sepsis in United States emergency departments. Critical Care Medicine, 35(8), 1928-36.

Wolff, M., & Bouadma, L. (2010). What procalcitonin brings to management of sepsis

in the ICU. Critical Care, 14(6), 1007. doi:10.1186/cc9330

World Health Organization. (2011). Health policy. Retrieved from http://www.who.int/gho/service_delivery/en/

Zhang, J., She, D., Feng, D., Yanhong, J., & Xie, L. (2011). Dynamic changes of serum soluble triggering receptor expressed on myeloid cells-1(sTREM-1) reflect sepsis severity and can predict prognosis: a prospective study. Biomedical Central Infectious Disease, 11(53), 1-7.

71

Appendix A: Nova IRB Exempt Letter

72

Appendix A: Organization IRB Exempt Letter

73

Appendix B: Letter of Project Support from Medical Surgical Unit

74

Appendix B: Letter of Project Support from Organization

75

Appendix C: Paper Copy of Severe Sepsis Tool

76

Appendix D: Proposed Severe Sepsis Algorithm

77

Appendix E: Sepsis Order Set in Medical Surgical Floor

78

Appendix E: Sepsis Order Set in Medical Surgical Floor

79

Appendix F: Education Flyer

Pilot study to employ a severe sepsis screening tool already utilized at Baptist Hospital of Miami and newly implement a severe sepsis algorithm for early identification of severe septic patients in a medical surgical non ICU unit, thus potentially reducing patient transfer to higher levels of care and hospital length of stays.

4 tower staff will be exposed to a 35-45 minutes educational seminar regarding the study and explanation of the developed sepsis algorithm. Please attend one of the below dates for the presentation.

Dates: January 5, 2015 11:00 am and 2:00pm January 6, 2015 11:00 am and 2:00pm January 7, 2015 11:00 am and 2:00pm January 8, 2015 07:30 am and 4:00am January 9, 2015 07:30 am and 4:00am January 12, 2015 07:30 am and 4:00am

Presenter Jorge Hirigoyen ARNP

80

Appendix G: International Classification of Diseases (ICD-9) Coding

995.90 - SIRS, NOS

995.93 - SIRS-NONINF W/O AC OR DS

995.91 - SEPSIS

995.92 - SEVERE SEPSIS

785.52 - SEPTIC SHOCK

81

Figure 1. Adaptation of structure, process, and outcome model (Donabedian, 1980).

Structures of Care Physical resources and

organizational resources

Processes of Care Implementation of severe

sepsis algorithm with biomarkers procalcitonin and lactic acid, guidelines

and order sets, and a proposed course of action

Health Outcomes Results: Decreased length

of hospital stay and transfer of patients to

higher level of care

82

Figure 2. Receiver operating curve (ROC).