Abstract of thesis entitled

“A Standardized Protocol to Increase Triage Accuracy for

Patients Presenting with Cardiac Symptoms in AED”

Submitted by

Choi Wing In

for the degree of Master of Nursing

at The University of Hong Kong

in August 2014

Cardiovascular disease is the disease involving the heart and the blood vessels. Clinical

manifestations included chest pain or chest discomfort, shortness of breath, palpitations,

dizziness etc. Cardiovascular diseases have significant mortality and morbidity. According to

the World Health Organization global burden of disease 2004 update, it is the leading cause

of death worldwide (WHO, 2008).

Chest pain is one of the typical signs of cardiovascular diseases, and it is also one of the

most commonly seen complaints among patients attending accident and emergency

departments (AED). The early identification of patients impending cardiovascular events is

particularly important, to allow prompt nursing actions and early interventions to be initiated.

However, in clinical practice, not all chest pain is related to myocardial ischemia. Chest

pain can be caused by many other conditions such as gastrointestinal, respiratory or

musculoskeletal problems, which are not cardiac in origin (Scher, 1995). It is important for

health care professionals in AED to differentiate life threatening cardiovascular events from

other conditions that require less urgent attentions. To ensure health resources are effectively

used and prioritize who are in need of immediate attentions, accurate triage decisions are

highly emphasized.

Currently there is no standardized assessment tool adopted for chest pain nursing triage in

accidents and emergency departments in Hong Kong, therefore a structured and standardized

guideline in performing assessments, applying interventions and triaging patients with chest

pain is needed.

In this dissertation, a translational research was conducted to aid developing a

standardized protocol to increase triage accuracy for patients presenting with cardiac

symptoms in accident and emergency department.

In the first chapter, the background and the need for a standardized protocol to aid chest

pain triage in AED will be introduced. In the second chapter, methods to search for relevant

evidence related to chest pain risk stratification in AED will be described. Critical appraisal

of the selected studies will be undertaken by using appraisal checklist of Scottish

Intercollegiate Guideline Network (SIGN). Several risk stratification tools have been

identified from the literature review, among those, the Front Door Score (FDS) was selected

to aid chest pain triage. In the third chapter, the implementation potential of the FDS will be

assessed, in terms of the target setting and audience, the transferability of the selected studies,

the feasibility to carry out the FDS and the cost & benefit ratio. The evidence based guideline

of using FDS in AED will also be discussed. In the fourth chapter, the implementation plan of

FDS will be discussed, including the communication plan, pilot study and evaluation plan of

the proposed guideline.

A Standardized Protocol to Increase Triage Accuracy for Patients

Presenting with Cardiac Symptoms in AED

by

Choi Wing In

BNurs. (Hon), HKU

A thesis submitted in partial fulfillment of the requirements for

the degree of Master of Nursing

at The University of Hong Kong

August 2014

i

Declaration

I declare that this thesis represents my own work, except where due acknowledgment is

made, and that it has not been previously included in a thesis, dissertation or report submitted

to this University or to any other institution for a degree, diploma or other qualifications.

Signed…………………………………………

CHOI Wing In

ii

Acknowledgments

I would like to express my gratitude to my dissertation supervisor, Ms. Idy Fu and Dr.

Elizabeth Hui, for their guidance and assistance with my dissertation throughout the process.

They have offered me inspirations and suggestions using their expert knowledge and research

experience.

I would also like to express my special thanks to Ms. Margaret Yuen, Nurse Specialist,

Accidents and Emergency Department, Pamela Youde Nethersole Eastern Hospital for

providing valuable data from the Accidents and Emergency Department.

iii

Contents

Declaration……………………………………………………………………………………………….i

Acknowledgments……………………………………………………………………………………….ii

Table of Contents…………………………………………………………………………………….iii-v

List of Appendices…………………………………………………………………………………..….vi

List of Abbreviations and Symbols…………………………………………………………………..vii

CHAPTER 1: INTRODUCTION

1.1 Background………………………………………………………………………….…...1

1.2 Significance …………………………………………………………………..………….2

1.3 Affirming the Need…………………………………………………………………….....3

1.4 Objectives………………………………………………………………………………...6

CHAPTER 2: CRITICAL APPRAISAL

2.1 Search Strategies

2.1.1 Keyword Search…………………………………………………………………......7

2.1.2 Inclusion and Exclusion Criteria…………………………………………………….8

2.2 Results………………………………………………………………………………........8

2.3 Appraisal Strategies……………………………………………………………............10

2.3.1 Relatively High level of evidence………………………………………………...10

2.3.2 Medium level of evidence……………………………………….……………….11

2.3.3 Low level of evidence …………………………………………………………..….12

2.4 Summary of Data

2.4.1 Purpose and objective………………………………………………………………13

2.4.2 Patient selection…………………………………………………………………….13

2.4.3 Risk stratification tools……………………………………………………………..13

iv

2.4.4 Outcome measures…………………………………………………………………14

2.4.5 Statistical Analysis…………………………………………………………………14

2.4.6 Findings…………………………………………………………………………….15

2.5 Synthesis of Data…………………………………………………………………….....16

2.5.1 GRACE score………………………………………………………………………17

2.5.2 TIMI risk score……………………………………………………………………..17

2.5.3 Front Door Score…………………………………………………………………...18

CHAPTER 3: TRANSLATION AND APPLICATION......................................................20

3.1 Transferability of the findings…………………………………………………………20

3.1.1 Target Setting……………………………………………………………………….20

3.1.2 Target Audience…………………………………………………………………….21

3.1.3 Philosophy of Care…………………………………………………………………22

3.1.4 Clients Benefitted from the Innovation…………………………………………….22

3.2 Feasibility of the Innovation

3.2.1 Freedom to Implement……………………………………………………………..23

3.2.2 Administrative Support……………………………………………………………..23

3.2.3 Staff Competence & Training………………………………………………………23

3.2.4 Staff Acceptability………………………………………………………………….24

3.2.5 Availability of Equipment, facilities & Resources…………………………………25

3.3 Cost-Benefit Ratio of the Innovation

3.3.1 Potential Benefits of the Innovation………………………………………………..25

3.3.2 Potential Risks of the Current Practice……………………………………………..26

3.3.3 Material Costs………………………………………………………………………26

3.3.4 Potential Non-material Costs and Benefits………………………………………...27

3.4 Evidence Based Practice Guideline ……………………………………………….29-31

v

CHAPTER 4: IMPLEMENTATION PLAN

4.1 Communication Plan

4.1.1 Identify Stakeholders……………………………………………………………….32

4.1.2 Potential Oppose……………………………………………………………………33

4.1.3 Process of Communicating the Plan………………………………………………..33

4.1.4 Setting Up a Communication Team………………………………………………..34

4.1.5 Staff Training……………………………………………………………………….35

4.2 Pilot Study

4.2.1 Objectives of the Pilot Study Plan……………………………………………….…36

4.2.2 Time Frame…………………………………………………………………………37

4.2.3 Patient Selection……………………………………………………………………37

4.2.4 Method……………………………………………………………………………...37

4.2.5 Evaluation of the Pilot Study……………………………………………………….38

4.3 Evaluation Plan

4.3.1 Identifying Outcomes………………………………………………………………39

4.3.2 When and How Often to take Measurement……………………………………….41

4.3.3 Data Analysis……………………………………………………………………….42

Conclusion………………………………………………………………………………..43

References……………………………………………………………………………44-46

Appendices…………………………………………………………………………..…47-63

vi

List of Appendices

Appendix 1: Search History…………………………………………………………………47

Appendix 2: Table of Evidence…………………………………………………………48-51

Appendix 3: SIGN Appraisal Checklists………………………………………………...52-59

Appendix 4: Level of Evidence and Grades of Recommendation by SIGN……………….60

Appendix 5: Costs of Implementation………………………………………………………61

Appendix 6: Front Door Score Data Entry Form………………………………….……62-63

vii

List of Abbreviations and Symbols

Abbreviations

ACS Acute coronary syndrome

AED Accidents and emergency department

AMI Acute myocardial infarction

CABG Coronary artery bypass graft

COS Chief of Service

CREC Clinical research ethics committee

DOM Department operating manager

ECG Electrocardiogram

FDS Front Door Score

GRACE the Global Registry of Acute Coronary Events

HCA Health care assistance

HFA/CS-ANZ the Heart Foundation of Australia and Cardiac Society of Australia and New

Zealand

NS Nurse Specialist

PCI Percutaneous coronary intervention

PURSUIT Platelet glycoprotein IIb/IIIa in Unstable angina: Receptor Suppression Using

Integrilin

ROC Receiver operating characteristic

STEMI ST-elevation myocardial infraction

TIMI Thrombolysis in myocardial infarction

WM Ward manager

Symbols

κ Kappa

1

CHAPTER 1

INTRODUCTION

1.1 BACKGROUND

Cardiovascular disease is the disease involving the heart and the blood vessels. Clinical

manifestations included chest pain or chest discomfort, shortness of breath, palpitations,

dizziness etc. Cardiovascular diseases have significant mortality and morbidity. According to

the World Health Organization global burden of disease 2004 update, it is the leading cause

of death worldwide (WHO, 2008).

Chest pain is one of the most common complaints among patients attending accident and

emergency departments (AED) (Ho et al., 2013). It can be a typical sign of coronary artery

disease. It is widely accepted that early interventions in these patients can reduce mortality

and morbidity (Speake et al., 2003). Therefore, it is important to stratify the risk of chest pain

accurately, provide prompt nursing actions and initiate early interventions for patients in AED,

with suspected coronary artery disease.

However, in clinical practice, not all chest pain is related to myocardial ischemia. Chest

pain can be caused by many other conditions such as gastrointestinal, respiratory or

musculoskeletal problems, which are not cardiac in origin (Scher, 1995). It is important for

health care professionals in AED to differentiate life threatening cardiovascular events from

other conditions. To ensure health resources are effectively used and prioritize who are in

2

need of immediate attentions, accurate triage decisions are highly emphasized. At present,

there are no standardized assessment tools to help nurses in accidents and emergency

departments in Hong Kong to stratifying the risk of chest pain in triage patients. Therefore

there is a need to implement a structured and standardized approach to aid nursing triage

decision making.

1.2 SIGNIFICANCE

According to the World Health Organization global burden of disease 2004 update,

cardiovascular diseases are the leading cause of death in the world, such diseases caused

almost 32% of all deaths in women and 27% in men in 2004 (WHO, 2008). The disease of

the heart is also the second leading cause of death in Hong Kong (Department of Health

annual report 2009/2010). Chest pain may be a signal of an impending life threatening

cardiac event, but it can also be caused by other non-cardiac conditions such as

gastrointestinal, respiratory or musculoskeletal problems, thus it is necessary to differentiate

the nature of chest pain, so as to identify who are in need of immediate treatment.

According to the American Heart Association (AHA), the door-to-first ECG time should

be less than ten minutes, and the door-to-balloon time should be within 90 minutes (AHA,

2013). That means every patient who attends AED presenting with chest pain should receive

first ECG within ten minutes, and for those who are having acute myocardial infarction

should receive percutaneous coronary intervention or thrombolytic treatment within 90

3

minutes. A revised concept of “first medical contact (FMC)-to-device time” is launched in

2013, representing both the recognition that the key issue is triaging and treating the patient

as soon as possible (AHA, 2013 STEMI guideline). Under AHA guidelines in managing

patients with chest pain, timely intervention is one of the major concerns. Accurate triage

decisions are crucial to allow prompt initial nursing actions and allow timely medical

interventions.

The purpose of triage is to assign patients’ urgency among all those who attend accidents

and emergency department. Decision making is a very crucial constituent in nursing triage.

Triage nurses are responsible to provide care for all patients who are pending medical

consultations. Also, triage nurses have a unique overview of the workload of the department,

and they play an important role in management of the patient flow through the department

(Chung, 2005).

1.3 AFFIRMING THE NEED

At present, patients attending accidents and emergency departments are firstly assessed

by triage nurses, assessments included medical history taking, interview on chief complaint

and vital signs taking. After triage assessments, patients are assigned with an acuity rating

based on a 5-level triage scale typically adopted by public hospitals in Hong Kong nowadays.

Patients are divided into 5 categories according to their medical conditions, category 1 to

category 5, which refer to critical, emergency, urgent, semi-urgent and non-urgent

4

respectively. For critical patients who are dying or in life-threatening condition, they are

accorded top priority and attended immediately by a team of medical and nursing staff

without waiting. According to the Triage Guideline published by Hospital Authority 2011

edition, for patients with suspected myocardial infarction or acute coronary syndrome with

ECG changes, category 1 (critical) is assigned. For patients with suspected unstable angina,

category 2 (emergency) is assigned. For patients with chest pain suspected to be of cardiac

origin with stable vital signs, category 3 (urgent) is assigned. For patients with suspected

non-cardiac chest pain, category 4 (semi-urgent) is assigned.

Triage decision is always a challenge to nurses as nurses are required to make quick

decisions when very limited understanding of patients’ problem and patient information are

available. Three possible outcomes on the acuity of triage decisions are: appropriate decision,

which means that the assigned triage category is optimal; under-triage, means the assigned

triage category is less acute than required; over-triage, which means the assigned triage

category is more acute than required. In the study of Goransson et al. (2005), there are

discrepancies between patient assessments made by physicians and nurses. It was found that

only 57.6% of the triage acuity ratings made by nurses were in line with that made by

physicians (Goransson et al., 2005).

Moreover, nurses tended to over-triage patients. The study of Chung (2005), had explored

the experiences of triage decision making of accidents and emergency nurses in Hong Kong.

5

She found that sometimes nurses were uncertain in triage decision making. Such uncertainties

have making them feel that triage decision makings were stressful and risky. Some of them

would choose to upgrade the triage category to prevent patient conditions from deteriorating.

Although triage guidelines are available, it can only be a reference for triage decision

makings. Many nurses reported that they followed the guideline loosely, as not all of the

situations related to patient conditions fitted well into the guideline categories (Chung, 2005).

In addition, some nurses may not be knowledgeable or experienced enough in performing

chest pain triage. As a result, under-triage, over-triage or discrepancy in triage decisions

among different nurses can occur, when nurses made triage decisions according to personal

clinical experiences or intuitions, but not based on standardized and valid assessment tools.

Inaccurate triage decisions can lead to inappropriate utilization of health resources and

adversed patient outcomes (Gerdtz & Bucknall, 1999). An incorrect triage category allocation

can lead to delay in treatment and death of a patient at worst. Such outcomes may have legal

consequences. An over-triage can induce inconsistency in triage decisions, and will invariably

lengthen the waiting time within the same triage category group. The trend to over-triage is

not a good practice in which the ultimate goal of triage cannot be achieved.

Furthermore, currently no standard guideline is specifically designed and adopted for

chest pain triage in accidents and emergency departments in Hong Kong. The American

Heart Association College of Cardiology had published guidelines on managing patients with

6

chest pain, but these guidelines are mainly focused on goals and algorithms in the

management of acute coronary syndrome or acute myocardial infarction, rather than on

nursing triage. A standardized approach for chest pain triage is currently not available.

1.4 OBJECTIVES

In view of the serious health consequences of coronary artery diseases which can be

presented initially by experiencing chest pain, and the importance of prompt actions in

managing these patients, accurate triage decision is crucial to ensure health resources are

effectively used in prioritized patients. Currently there is no standardized assessment tool

adopted for chest pain nursing triage in accidents and emergency departments in Hong Kong,

therefore a structured and standardized guideline in performing assessments, applying

interventions and triaging patients with chest pain is needed.

The objectives of this study are:

1. To identify a valid and reliable risk stratification tool from existing literatures to aid chest

pain triage in emergency department.

2. To develop a standardized protocol for chest pain triage in emergency department.

3. To assess the transferability and feasibility of the proposed chest pain triage protocol.

4. To design an implementation and evaluation plan for the chest pain triage protocol.

7

CHAPTER 2

CRITICAL APPRAISAL

In this chapter, method to search for relevant evidence related to chest pain risk

stratification in accidents and emergency departments will be described. Systematic search on

different database using keywords search will be demonstrated. Screening on the titles and

abstracts were done using inclusion and exclusion criteria to identify relevant research

evidence. Critical appraisal of evidence was done by using the appraisal checklist of Scottish

Intercollegiate Guidelines Network (SIGN). Research articles will be listed as high, medium

or low level of evidence. Summary and synthesis of data will also be provided.

2.1 SEARCH STRATEGIES

2.1.1 Keyword search

Electronic database including PubMed, Ovid MEDLINE, CINAHL were used for the

systematic search. The keywords search comprise of three parts, firstly keywords of: “chest

pain”, “acute coronary syndrome” and “angina” were searched individually and combined for

“OR” search. Secondly, “risk stratification” and “risk score” were input and also searched as

“OR”. Thirdly, “emergency department” was searched. Lastly, all the three advanced searches

were combined for “AND” search. Keywords of “computer tomography”, “computed

tomography” and “troponin” were added to the advanced search for “NOT” search. As a

result, there were 188 results from PubMed, 157 results from Ovid MEDLINE and 43 results

8

from CINAHL.

2.1.2 Inclusion and Exclusion Criteria

The articles were selected if: (1) its target population are adult patients attending AED

with chest pain as their chief compliant, (2) the risk stratification method or assessment tool

was examined.

Articles were excluded if: (1) the studied population was limited to some special

population for examples: patients with renal disease or cocaine user, and (2) the risk

stratification method studied were not nursing related, for examples the study of

echocardiography or computer tomography.

2.2 RESULTS

Using the inclusion and exclusion criteria, the titles and abstracts of the total 388 articles

were screened. Articles were limited to English language and full text. Full text was also

screened for potential eligible articles. 43 articles were identified in CINAHL, 41 were

excluded and 2 were selected. 257 articles were identified in Ovid MEDLINE, 150 were

excluded and 7 were selected. 188 articles were identified in PubMed, 179 were excluded and

9 were selected. Duplicated and irrelevant studies were eliminated. The reference lists from

these selected articles were also screened to identify any relevant literatures.

Finally, eight studies were selected for the review, six studies were identified from

PubMed, one study was identified from CINAHL and one study was identified from Ovid

9

MEDLINE (see Appendix 1). The useful data from the studies was extracted and summarized

in the “Table of Evidence”. The columns in the Table of Evidence were listed according to

bibliographic citation, study type, sample size, patients’ characteristics, assessment tool,

comparison, outcome measures and conclusion (see Appendix 2). No RCTs were available in

the systematic search. In these eight articles, five were prospective observational cohort

studies (Chase et al., 2006; Pollack et al., 2006; Ramsay et al., 2007; Cullen et al., 2013; Lee

et al., 2011) and three were descriptive studies (Ho et al., 2013; Morris et al., 2006; Lyon et

al., 2007).

The descriptive study done by Ho et al. (2013) was conducted in Hong Kong. Three

studies were conducted in the United Kingdom (Morris et al., 2006; Ramsay et al., 2007;

Lyon et al., 2007). Three studies were conducted in the US (Chase et al., 2006; Pollack et al.,

2006; Lee et al., 2011), and one study was conducted in Australia (Cullen et al., 2013).

Among the eight studies included in this review, two of the observational cohort studies

aimed to validate the TIMI risk score (Chase et al.,2006; Pollack et al., 2006); two descriptive

studies aimed to evaluate the TIMI risk score and Front Door Score (Ho et al., 2013; Morris

et al., 2006); two studies compared the TIMI risk score and GRACE score on their ability to

risk stratifying chest pain patients in emergency departments (Ramsay et al., 2007; Lyon et al.,

2007); and finally two studies compared TIMI risk scores with GRACE score, PURSUIT

score and HFA/CS-ANZ guideline (Cullen et al., 2013; Lee et al., 2011).

10

2.3 APPRAISAL STRATEGIES

The critical appraisal checklist designed for diagnostic studies of Scottish Intercollegiate

Guidelines Network (SIGN, 2006) was used to assess the quality of the eight selected articles

(see Appendix 3). In the SIGN grading system, the level of evidence is ranked into four

categories, listed from level one to level four (see Appendix 4). Level one is the highest level

of evidence including high quality meta-analyses, systematic reviews of RCTs or RCTs.

Level two includes case control studies and cohort studies. Level three includes non-analytic

studies such as case reports or case series. Level four is expert opinions. Studies ranked level

one or two will be rated as “++” if it is of high quality with very low risk of confounding or

bias; rated as “+” if it is well-conducted with low risk of confounding of bias; rated as “-“ if it

is of a high risk of confounding or bias.

2.3.1 Relatively High level of evidence

The prospective observational cohort studies done by Chase et al. (2006), Cullen et al.

(2013) and Lee et al. (2011) attained high level of evidence, and were rated as “2++”,

according to the critical appraisal checklist of Scottish Intercollegiate Guidelines Network

(SIGN, 2006).

All three studies have clearly stated the study objectives and designs. The study of Chase

et al. (2006) aimed to validate the use of TIMI risk score in unselected chest pain patients in

emergency departments, with the diagnostic accuracy of TIMI risk score being examined; the

11

study of Cullen et al. (2013), compared the diagnostic accuracy of HFA/CS-ANZ guideline,

GRACE score and TIMI risk score on 30-day adverse cardiac outcomes; while the study of

Lee et al. (2011) compared the diagnostic accuracy of PURSUIT score, GRACE score and

TIMI risk score. All three studies had large sample size enrolled without inappropriate

exclusion criteria. Sample size ranged from 948 to 4743. Consecutive sequences of samples

were enrolled. Detailed descriptions and definitions of outcome measures including

diagnostic criteria for ACS, all-cause mortality, AMI, revascularization and unstable angina

were clearly stated and defined in all three studies. In the study of Chase et al. (2006), method

to blinding of researchers was mentioned, the investigators were unaware of the calculated

score, and this could minimize bias when evaluating the patients’ outcomes. Precise statistical

results for diagnostic accuracy statistics on sensitivity, specificity, and the relative risk of

outcomes for each score component were reported in all three studies, provide with

confidence intervals.

2.3.2 Medium level of evidence

Two descriptive studies (Ho et al., 2013; Morris et al., 2006) and two prospective

observation cohort study (Pollack et al., 2006; Ramsay et al., 2007) in this review attained

medium level of evidence, and were rated as “2+”.

Aims and objectives of the four studies were clearly stated. The study of Ho et al. (2013)

aimed to evaluate the ability of Front Door Score to enhance triage accuracy. Morris et al.

12

(2006) validated TIMI risk score and Front Door Score in undifferentiated chest pain

population in emergency departments. Pollack et al. (2006) and Ramsay et al. (2007)

examined TIMI risk score and GRACE score. The sample selection methods were

appropriate with consecutive eligible samples being enrolled. Outcome measures were clearly

stated in all four studies. Some bias and confounding factors was observed in the study design.

For examples, in the study of Ho et al. (2013), foreigners were excluded in the study due to

possible language barriers, there is possible selection bias. Besides, methods to minimize bias

of researchers in evaluating patient outcomes were not mentioned in these four studies, thus

bias may occur upon data collection and outcome evaluation process.

2.3.3 Low level of evidence

The descriptive study of Lyon et al. (2007) attained low level of evidence of “2-“,

according to the critical appraisal checklist of Scottish Intercollegiate Guidelines Network

(SIGN, 2006). Some limitations of the study design can be identified. Firstly, the method of

sample selection, the inclusion and exclusion criteria for sample selection was not mentioned.

Both GRACE and TIMI risk scores were calculated retrospectively, instead of prospectively.

Besides, definitions of outcomes measures were not clearly specified. Also, 24% of recruited

eligible sample were excluded due to incompleteness of data to calculate the score. Methods

of blinding of researchers were also not addressed.

13

2.4 SUMMARY OF DATA

2.4.1 Purpose and objective

All eight articles have clearly stated purpose and objectives. Seven of them aimed to

validate one or more risk stratification tools to risk stratify unselected patients attending AED

for chest pain (Morris et al., 2006; Chase et al., 2006; Pollack et al.,2006; Ramsay et al., 2007;

Lyon et al., 2007; Cullen et al., 2013; Lee et al.,2011). The diagnostic accuracy of the risk

stratification tools was compared. One aimed to examine the effectiveness of using the risk

stratification tools to increase triage accuracy (Ho et al., 2013).

2.4.2 Patient selection

All samples enrolled for the eight studies were adult patients attending emergency

departments with the chief compliant of chest pain. Those patients with clear signs of

non-cardiac chest pain, for example due to trauma or surgery, were excluded in these studies.

Sample size of these eight studies ranged from 200 to 4743.

2.4.3 Risk stratification tools

All of the eight studies examined risk stratification tools that are designed to risk stratify

chest pain patients. Their relationships with adverse cardiac outcomes were studied in seven

studies, and the relationship of TIMI risk score and Front door score with triage accuracy was

studied in one study. TIMI risk score was validated in two of the studies (Chase et al., 2006;

Pollack et al., 2006). In the studies of Ho et al. (2013) and Morris et al. (2006), both TIMI

14

risk score and Front Door Score were examined. TIMI risk score were evaluated and

compared with GRACE score in the studies of Ramsay et al. (2007) and Lyon et al. (2007).

The study of Cullen et al. (2013) compared the TIMI risk score with GRACE score and

HFA/CS-ANZ guideline, while that of Lee et al. (2011) compared TIMI risk score with

GRACE score and PURSUIT score.

2.4.4 Outcome measures

All eight studies have similar outcome measures. Outcome measures in four studies were

all-cause mortality, AMI, and revascularization by percutaneous coronary intervention or

coronary artery bypass surgery (Chase et al., 2006; Pollack et al., 2006; Ramsay et al., 2007;

Lee et al., 2011). Outcome measures in three studies included all-cause mortality, AMI,

revascularization by percutaneous coronary intervention or coronary artery bypass surgery

and acute coronary syndrome (Morris et al., 2006; Lyon et al., 2007; Cullen et al., 2013). In

Ho et al. (2013) study, the outcome measure was the triage reliability using TIMI risk score,

Front Door Score and 5-level triage scale when compared to the gold standard, which is the

final physician diagnoses.

2.4.5 Statistical Analysis

In the study of Chase et al. (2006) and Pollack et al. (2006), the reliability of TIMI risk

score were illustrated by its correlation with the incidence rate of AMI, death or

revascularization within 30-days in patients of each score group. The analyses were done by

15

using chi-square testing and Cochran-Armitage trend test. In another five studies (Morris et

al., 2006; Lyon et al., 2007; Cullen et al., 2013; Ramsay et al., 2007; Lee et al., 2011), the

reliability of the risk stratification tools including TIMI risk score, Front Door Score,

GRACE score, HFA/CS-ANZ guideline and PURSUIT score were illustrated by their

sensitivity and specificity. By using the sensitivity and specificity of each tool, Receiver

operating characteristic (ROC) curves were plotted. The diagnostic accuracy of different tools

were compared by the area under the ROC curves. In the study of Ho et al. (2013), the triage

reliability of nurses using TIMI risk score, Front Door Score and 5-level triage scale were

compared, by their agreement rates with the gold standard, analyzed by kappa statistics.

2.4.6 Findings

Among the eight articles included in this review, seven have shown that TIMI risk score

was correlated with adverse cardiac outcomes (Morris et al., 2006; Chase et al., 2006; Pollack

et al., 2006; Ramsay et al., 2007; Lyon et al., 2007; Cullen et al., 2013; Lee et al., 2011). Its

correlation to the likelihood of adverse outcomes was demonstrated by a p-value of < 0.001.

In two studies, Front Door Score was also found to be correlated with adverse cardiac

outcomes (Morris et al., 2006; Ho et al., 2013). Its diagnostic accuracy was shown to be

inferior when compared to TIMI risk score, but its clinical importance was addressed. Front

Door Score was also proved to be effective to enhance reliability of chest pain triage (Ho et

al., 2013).

16

Diagnostic accuracy was represented by area under the Receiver operating characteristic

(ROC) curves. By calculating the sensitivity and specificity of the tools, ROC curves can be

plotted and diagnostic accuracies can be obtained. For a diagnostic accuracy equals to one,

mean that the risk stratification tool is of 100% accuracy. The calculated diagnostic accuracy

of GRACE score ranged from 0.73(Lee et al., 2011) to 0.82(Ramsay et al., 2007); that for

TIMI risk score ranged from 0.74(Ramsay et al., 2007) to 0.79(Lyon et al., 2007); that for

Front Door Score was 0.70 (Morris et al., 2006). Among the four studies which compared the

risk stratification tools (Ramsay et al., 2007; Lyon et al., 2007; Cullen et al., 2013; Lee et al.,

2011), three of them (Ramsay et al., 2007; Lyon et al., 2007; Cullen et al., 2013) have shown

that GRACE score is slightly more superior to TIMI risk score. In the study of Lee et al.

(2011), results have shown that TIMI risk score performs better. To conclude, the diagnostic

accuracies of TIMI risk score, Front Door Score & GRACE score were all found to be

satisfactory (Ramsay et al., 2007; Lyon et al., 2007; Cullen et al., 2013; Lee et al., 2011).

2.5 SYNTHESIS OF DATA

Several risk stratification tools were developed for the use to risk stratify chest pain

patients in emergency departments. Among those, Global Registry of Acute Coronary Events

(GRACE) score and Thrombolysis in myocardial infarction (TIMI) risk score are the most

widely validated tools.

17

2.5.1 GRACE score

GRACE score consists of 8 variables: 1)age, 2)heart rate, 3) SBP, 4)Killip class,

5)Creatinine, 6) cardiac arrest at admission, 7)ST segment deviation, 8) elevated cardiac

enzyme. The calculation of GRACE score has been published as an online calculator

(http://www.outcomes-umassmed.org/Grace). By imputing patient data via the online

calculator, risk of death and risk of AMI at different time intervals will be calculated. Thus it

can be used as a reference to guide clinical management according to individual’s risk. It can

also be used in emergency departments which focus is on early and accurate identification of

ACS or AMI.

2.5.2 TIMI risk score

TIMI risk score consists of 7 predictors, each score one point. The predictors are: 1)

age>65, 2) known coronary heart disease, 3) at least 3 risk factors for coronary heart disease,

4) aspirin use in previous 7 days, 5) at least 2 angina episodes in last 24 hours, 6) ST segment

deviation of 0.5mm 7) elevated cardiac markers. TIMI risk score ranges from zero to seven.

Calculation of TIMI risk score of chest pain patients in emergency departments helps to

identify who are at higher risk for ACS or AMI. Patients with score of zero are considered

low risk. Patients score one to four are considered intermediate risk, and those who scored

five to seven are considered high risk. Such score can guide clinical decision making. Unlike

GRACE score, which is derived from ACS populations, TIMI risk score has been proved to

18

perform well in unselected chest pain population in emergency departments.

2.5.3 Front Door Score

Front Door Score (FDS) is a simplified version of TIMI risk score, with the exclusion of

cardiac markers component. Despite the sensitivity and specificity of FDS is not as good as

the full score, it retain its ability to risk stratify the chest pain population in emergency

departments.

The ability of GRACE score, TIMI risk score and Front Door Score to risk stratify chest

pain patients in accidents and emergency departments were validated in various studies. All

of them are of good diagnostic accuracy. However, the complexity of GRACE score has

limited its usage in accidents and emergency departments to stratify risk of adverse outcomes

in chest pain patients. For examples, an online calculator is needed to obtain the GRACE

score, which can be a time-consuming procedure. Also, some components to calculate

GRACE score are difficult to be obtained at triage level, making it not feasible for triage

nurse to use it to aid chest pain triage. For TIMI risk score, despite its ability to predict risk of

adverse cardiac outcomes in chest pain patients, its usage in nursing triage is also limited. It is

because cardiac marker component which is needed for calculation of TIMI risk score cannot

be initially obtained at the time of triage. For Front Door score, it is also of good diagnostic

accuracy, despite its exclusion of the cardiac marker component in its calculation. It is also

19

validated in various studies for its sensitivity & specificity. In contrast to GRACE score or

TIMI risk score, Front Door Score is easy to calculate, which consists of only six components

in its calculation. All six components can be easily obtained during history taking process at

triage level. It is practical and feasible for triage use where cardiac markers cannot be initially

obtained and thus can be used to inform triage decision making.

In view of no standardized assessment tool is adopted in Hong Kong to help nurses in

accidents and emergency departments to stratify the risk of chest pain in triage patients, Front

Door Score can be consider as a standard tool to enhance chest pain triage decisions.

To conclude, patients attending emergency departments for chest pain should be firstly

assessed by a trained triage nurse, using a standardized assessment protocol. The Front Door

Score is proposed to be adopted as the standardized guideline. Front Door Score has been

shown to be a reliable tool to enhance chest pain triage accuracy, predict the risk of

developing adverse cardiac outcomes and thus ensure effective use of health resources in

prioritized patients.

20

CHAPTER 3

TRANSLATION AND APPLICATION

CHAPTER SUMMARY

As discussed in the previous chapter, cardiovascular diseases have significant mortality

and morbidity. Chest pain is one of its typical signs, and it is also one of the most commonly

seen complaints among patients attending accident and emergency departments. The early

identification of patients impending cardiovascular events is particularly important, to allow

prompt nursing actions and early interventions to be initiated. Several risk stratification tools

have been identified from a literature review to risk stratify chest pain patients in accident

and emergency departments. The Front Door Score (FDS) was selected to aid chest pain

triage.

In this chapter, the implementation potential of the FDS will be assessed, in terms of the

target setting and audience, the transferability of the selected studies, the feasibility to carry

out the new innovation and the cost & benefit ratio. The evidence-based guideline of using

the FDS to aid chest pain triage in AED will also be discussed.

3.1 TRANSFERABILITY OF THE FINDINGS

3.1.1 Target Setting

The target setting for the implementation of FDS will be the accident and emergency

department (AED) in a public hospital of the Hong Kong East cluster. It is an acute hospital

21

with 1597 beds and 24-hour AED service.

The eight studies selected for the review were all conducted in AEDs. One study was

conducted in Hong Kong (Ho et al., 2013), three in United Kingdom (Morris et al., 2006;

Ramsay et al., 2007; Lyon et al., 2007), three in the US (Chase et al., 2006; Pollack et al.,

2006; Lee et al., 2011) and one in Australia (Cullen et al., 2013). All AEDs from different

countries in the selected studies are 24-hours operating, providing acute care to patients who

present without prior appointments. Patients presenting with chest pain symptoms with

possible acute coronary syndrome accounts for a significant proportion of the overall

attendance of the AEDs in the selected studies, this share similar conditions to the target

setting in this proposal.

3.1.2 Target Audience

The target audience for the new innovation is patients attending the AED, aged 18

year-old or above, with chest pain as chief compliant. Those patients who are below age of 18

years; mentally incapability for communicate e.g. post stroke; have clear signs of non-cardiac

chest pain, for example due to trauma or injury; critically ill or having hemodynamically

unstable vital sign on presentation and initial triage ECG already showed obvious ST

segment elevation myocardial infarction (STEMI) will be excluded.

All participants in the eight selected studies were aged 18 years-old or above, with chest

pain as their chief complaint. Although the ethnicity of the participants in the studies may not

22

be the same as the target audience, both groups are unselected populations in the AEDs, the

participants are still comparable to our proposed target audience.

3.1.3 Philosophy of Care

The target hospital was established in 1993, as a major acute hospital providing a full

range of specialist services to residents of the eastern district of Hong Kong Island. The

hospital has been upholding the philosophy of providing holistic people-centered quality care

through love, dedication and teamwork. The AED of the target hospital is committed to

deliver timely, quality emergency services to our patients.

As such philosophy is in-line with the aim of the proposed innovation. The aim of

adopting FDS is to risk stratify chest pain patients in AED, identify their risk of developing

significant cardiac events and inform triage decisions, thus optimize treatments and improve

patient outcomes. The innovation can ensure the high quality emergency services.

3.1.4 Clients Benefitted from the Innovation

All those patients presenting to the AED with undifferentiated chest pain will be

benefitted from the innovation. According to the statistical report of the target AED setting in

2012, the annual attendance of patients in 2012 was 146,000 (HA statistical report, 2012).

Among them, 10% were admitted with chest pain as their chief compliant. The proposed

innovation targeted this group of patients. By the calculation based on 2012 statistics, it is

estimated that 14,600 patients can benefit from the innovation annually.

23

3.2 FEASIBILITY OF THE INNOVATION

3.2.1 Freedom to Implement

The new innovation can be commenced with the approval from the administrative and

management team in the proposed setting. They will be involved and consulted for

implementation of the evidence-based guideline. Meanwhile, evaluation plans, cost & benefit

analysis of the innovation will also be presented to them before the commencement. The

guideline will also be proposed to the Clinical Research Ethics Committee (CREC) of the

corresponding cluster. With both approvals, the guideline can be formal launched.

3.2.2 Administrative Support

Evidence-based practice is highly emphasized in the proposed setting. There are existing

nurse development programs ongoing in the department, including sharing sessions for nurses

to share ideas and new research findings on emergency nursing care. Staffs are also

enthusiastic to update themselves with new knowledge, through attending various trainings

and workshops. These have shown that the organization climate is conducive to research

utilization. As both frontline and management staffs are very supportive to new innovations,

the new evidence-based guideline will be highly appreciated.

3.2.3 Staff Competence & Training

Our nurses are competent and adequately trained for triage skills and knowledge. All

nurses with two years or more AED working experience were sent for triage training

24

workshops. They are qualified to perform triage after passing the triage audit. Updated triage

guideline published by the Hospital Authority is also readily available for all nurses in the

department. Triage nurses are being regularly audited by senior nurses about their triage skills

and competence.

To implement the FDS as a new innovation in the department, introduction of the FDS to

all nursing and medical staff is needed. Therefore training sessions will be offered by the FDS

communication team to all staff to introduce the philosophy of care of the FDS, communicate

corresponding research evidence and discuss the new triage workflow. The target population

for applying FDS, its calculation, and the subsequent nursing actions for patients of each

score group will also be discussed.

3.2.4 Staff Acceptability

To incorporate FDS in current triage practice, extra efforts will need to be made by triage

nurses. Due to the heavy workload and fast working pace at triage station, nurses may find it

difficult to adopt the change.

In order to facilitate the usage of FDS, orientation of the FDS and the benefits of adopting

it must be addressed to management team, frontline nurses and the doctors in order to gain

their support. We need to emphasize on the possible positive outcomes bring out by FDS to

aid chest pain triage, such as increase nursing triage accuracy, improved patient outcomes,

improved nursing autonomy, improved staff satisfaction and reduced health costs etc.

25

During the implementation process, feedbacks and comments from staff will be taken into

account for the evaluation of FDS. Modification and fine-tuning of the FDS guideline can be

made.

3.2.5 Availability of Equipment, facilities & Resources

In the proposed setting, well-equipped meeting rooms are readily available for conducting

the training sessions. Necessary equipment like computers and projectors were installed and

ready to be used.

The FDS is an assessment tool to risk stratifies the risk of developing adverse cardiac

outcomes in chest pain population to aid chest pain triage. Neither license nor copyright

charge is needed, no extra equipment is needed for its usage in triage. To incorporate the FDS

into current triage practice, simple flow chart will be developed for nurses’ quick reference at

triage station.

3.3 COST-BENEFIT RATIO OF THE INNOVATION

3.3.1 Potential Benefits of the Innovation

The FDS is a simplified version of TIMI risk score, which was originally designed to risk

stratify chest pain population in AED. By calculating the FDS, patients can be identified as

high, intermediate or low risk in terms of their susceptibility in developing adverse cardiac

outcomes. Integrating FDS into current triage system allows a more accurate, systematic and

specific assessment for patients with undifferentiated chest pain. In short term, proper

26

allocation of resources can be made possible, early interventions can be initiated. In long term,

cost from mortality and morbidity in patients impending coronary heart diseases can be

reduced from prompt & timely intervention after more accurate triage.

3.3.2 Potential Risks of the Current Practice

Currently no standardized guideline is in use for chest pain triage in the proposed setting.

The American Heart Association (AHA) resuscitation guidelines 2010 guidelines mainly

focused on goals and algorithms in the management of ST-elevation myocardial infarction,

rather than on nursing triage. Although HA triage guideline is available, many nurses found

that it can only be a reference for triage decision making. The study of Chung (2005) has

found that many nurses were not confident in chest pain triage, tendencies to over-triage

patients often occurred, ultimate goal of triage was therefore not achieved. Besides, the study

of Ho (2013) demonstrated that the accuracies of chest pain triage of nurses using HA triage

guideline are far from satisfactory. Subsequently, such inaccurate triage decisions can lead to

inappropriate utilization of resources and adverse patient outcomes. Therefore, the accuracies

of chest pain triage in the proposed setting cannot be improved if the current practice is not

modified.

3.3.3 Material Costs

The material costs for implementing the FDS in the proposed setting includes the cost for

the printing of the reference FDS flow charts and FDS forms. For the meeting rooms and

27

necessary equipment like computers and projectors are readily available in the proposed

setting, therefore no extra expenditure in this area is needed.

Training sessions on FDS will be conducted during regular nursing training sessions held

weekly in the department, which will last for 1 hour per session. All nurses will be arranged

to attend. In the proposed AED setting, there are 68 nurses in total, the man-hour required

will be 68. The cost of implementation is summarized in Appendix 5.

3.3.4 Potential Non-material Costs and Benefits

The implementation of FDS into current triage practice may need extra effort make by

triage nurses. This may increase the workload and stress of our nurses. Nurses’ morale may

temporarily reduce due to the frustration arise from the innovation. A proper training and staff

support is therefore necessary. The appreciation of staff contributions to the usage of FDS

should be highlighted. Evaluation of the innovation is also crucial to fine-tune the guideline

and allow staff to express their concerns and difficulties being encountered, which could help

to reduce staff frustration.

The FDS can help to aid chest pain triage and increase triage accuracy, which can in turn

increase staff confidence and job satisfaction. In long term, improved patient outcomes are

also expected.

To conclude, the aim of the innovation is to maximize patients’ benefits and minimize

potential risks and costs. The FDS was reported to be effective in improving patient outcomes,

28

and the costs of implementation are minimal. The potential benefits of FDS have outweighed

the potential costs. The findings from the reviewed studies are considered transferrable and

the FDS is feasible to be implemented in the proposed setting.

29

3.4 EVIDENCE-BASED PRACTICE GUIDELINE

Title of Guideline

A standardized protocol to increase triage accuracy for patients presenting with cardiac

symptoms in AED

Purpose of the Guideline

The purpose of this guideline is to help nurses to identify the risk of developing adverse

cardiac outcomes in chest pain patients, and guide triage decision makings. This will in turn

result in improved chest pain triage accuracy.

Objectives of the Guideline

1) To provide a practical and systematic approach to perform chest pain triage

2) To stratify the risk in patients with undifferentiated chest pain and sort out those high risk

group for early assessment & intervention

3) To improve triage accuracy for patients presenting with undifferentiated chest pain in

AED

Target Audience

Inclusions:

Age of 18year-old or above

Present to AED with chest pain as chief compliant

Stable vital signs at presentation

30

Exclusions:

Age below 18 year-old

Mentally incapable of communication

Obvious non-cardiac origin account for the chest pain e.g. trauma or injury.

Critically ill or hemodynamically unstable at presentation

Initial triage ECG showed obvious STEMI

Recommendations

The guideline is developed based on the eight reviewed articles (see appendix 2). The

evidence of these reviewed articles are graded according to the Level of Evidence (SIGN,

2011), and the recommendations suggested are graded according to the Grade of

Recommendation (see appendix 4).

Recommendation 1.0 B

FDS should be used in AED for risk stratification as well as prediction of outcome.

(2+ Morris et al, 2006; 2++ Chase et al, 2006; 2+ Pollack et al, 2006; 2+ Ramsay et al, 2007;

2- Lyon et al, 2007; 2++ Cullen et al, 2013; 2++ Lee et al, 2011)

Evidence: FDS was found to be positively correlated with adverse cardiac outcomes, in terms

of higher score with higher risk (Morris et al., 2006; Chase et al., 2006; Pollack et al., 2006;

Ramsay et al., 2007; Lyon et al., 2007; Cullen et al., 2013; Lee et al., 2011). Its diagnostic

31

accuracy was also found to be satisfactory (Ramsay et al, 2007; Lyon et al, 2007; Cullen et al,

2013; Lee et al, 2011)

Recommendation 2.0 C

Triage nurses should integrate FDS with clinical judgment and current 5-level triage

scale when performing chest pain triage.

(2+ Ho et al, 2013)

Evidence: FDS can be used in combination with clinical judgment and be incorporated to

5-level triage scale. In which resulting in improvement in patient documentation, assessment,

and interventions, as well as alleviates the uncertainties in performing chest pain triage (Ho et

al, 2013).

Recommendation 3.0 B

Nurses should perform triage by obtaining the circumstantial information and vital

signs, and then perform FDS for eligible patients.

(2+ Ho et al, 2013; 2++ Chase et al, 2006; 2+ Pollack et al, 2006)

Evidence: FDS cannot be used in isolation (Chase et al, 2006; Pollack et al, 2006). The

likelihood of an adverse outcome in patients with a FDS score of zero is not zero, FDS should

therefore be used in conjunction with clinical judgment (Ho et al, 2013; Pollack et al, 2006).

32

CHAPTER 4

IMPLEMENTATION PLAN

CHAPTER SUMMARY

In the previous chapter, the implementation potential of Front Door Score (FDS) in terms

of target audience, target setting, transferability of the selected studies and feasibility of the

FDS has been assessed. Based on the translation and application of the evidence, an evidence

based guideline of integrating FDS into current triage practice has been developed. In this

chapter, the implementation plan of FDS will be discussed, including the communication plan,

pilot study and evaluation plan of the proposed guideline.

4.1 COMMUNICATION PLAN

4.1.1 Identify Stakeholders

To integrate FDS into current triage practice, one important step is to identify all the

stakeholders involved. Anyone who is affected by the FDS is considered the stakeholders.

The stakeholders of the proposed guideline include the Chief of Service (COS), Department

Operating Manager (DOM), ward managers (WM), Nurse Specialist (NS), all nursing staff,

doctors and health care assistances (HCA) within the proposed AED setting, as well as all

targeted subjects of the proposed guideline.

Nurses who are eligible to perform triage are primarily important as they are involved in

the assessment and calculation of FDS as well as divert patient for subsequent management

33

based on the risk stratification.

HCAs are involved in performing ECG, their availability and willingness to quickly

perform the ECG is crucial in order to shorten the door-to-ECG time.

The role of doctors in the implementation of FDS is to interpret the first triage ECG and

identify those with ST-segment elevation myocardial infract (STEMI) for immediate

interventions. Secondly the attending doctors help to manage the patients and make the

provisional diagnosis accordingly.

Lastly are the targeted subjects, they are the patients attending the AED with

undifferentiated chest pain.

4.1.2 Potential Oppose

The potential opposes for the proposed guideline will be the nursing staff, doctors as well

as the target subjects. However, the proposed guideline is incorporated into current triage

workflow except with the additional FDS calculation. It imposes minimal burden to current

triage system as well as minimal increase in workload of triage nurses, the opposition from

staff is expected to be limited. The proposed guideline also imposes minimal risk and delay to

target subjects and therefore the opposition from target subjects also expected to be small.

4.1.3 Process of Communicating the Plan

In order to obtain support and approval from the administrators, the NS will be the first

one to be consulted as she is experienced and she also has frequent communication with the

34

administrators. The details of the FDS guidelines will be discussed with her. Later on, the NS

will be invited to communicate with COS, DOM & WM. Further communication with the

administrators on the evidence of the proposed guideline will be arranged. The discussion

includes the evidence of the proposed guideline, its practical feasibility and logistic

arrangement. Support from these administrative staff is crucial to the implementation of the

new guideline.

After getting the support from the departmental administrative staff, the new guideline

should be proposed to the Clinical Research Ethics Committee (CREC) of the corresponding

cluster, before it can be launched. As mentioned in Chapter 2, the proposed guideline fulfilled

the criteria of patient safety, reasonable prospects of improving healthcare management,

favorable risk-benefit ratio, foreseeable minimal risks and adverse outcomes, therefore the

chance for success in obtaining approval from CREC is high.

4.1.4 Setting Up a Communication Team

After the approval for implementation is obtained from CREC and departmental

administrative staff, communication process to all staff will begin. The objective of the

communication plan is to ensure that all staff involved in the new guideline to understand and

follow the FDS workflow.

A communication team will be formed to facilitate the implementation of the guideline.

The communication will include the NS, three experienced nurses, one doctor and the

35

proposer. The communication team promotes the new guideline to all AED nurses and

doctors, shares the evidence regarding FDS, conducts staff training, monitors implementation

process and conducts evaluation.

4.1.5 Staff Training

AED staff including all doctors and nurses will be involved in the staff training. The

evidence based guideline of integrating FDS into current triage practice will be introduced.

The communication team will share the evidence from the literatures, regarding the benefits

of using FDS in chest pain triage, the limitations of current chest pain triage practice, and

possible changes that the proposed guideline might bring out. Besides, the practical feasibility

and logistics arrangement of FDS will also be discussed.

Based on the shift duty of AED nurses, the most appropriate time to disseminate

departmental message is during the daily meetings held every afternoon. The communication

team will make use of such time to introduce the protocol. Meanwhile, training sessions on

the implementation of FDS guideline will also be conducted during regular nursing training

sessions which are held weekly in the department. All nurses will be arranged to attend.

Compulsory departmental meetings are held monthly in AED for AED doctors, briefing

sessions of FDS can therefore be taken place in the meetings. By estimation, in three monthly

meetings, all AED doctors will be acknowledged about the new FDS guideline. Details of the

proposed guideline will also be announced via staff intranet emails.

36

The target time frame for the dissemination of the proposed guideline among AED staff is

within 3 months. Within the three months, all nurses are expected to have attended the

weekly nursing training sessions at least once, and all doctors are expected to be

acknowledged with the proposed guideline. Once the initial dissemination process finished,

then a guardian program lead by the communication team will proceed. At least one nurse of

the communication team will be arranged to work in the triage station in each shift, to guide

the triage nurses and ensure that they manage to follow the proposed guideline correctly. To

ease the use of FDS, a hard copy of the evidence based guideline and the new chest pain

triage workflow will be kept at triage station. A quick reference showing eligible subjects

criteria and FDS calculation will also be available.

4.2 PILOT STUDY

Before the formal launch of the FDS guideline, a pilot study will be conducted. The

purposes of the pilot study include evaluating the effectiveness of the FDS, testing out the

feasibility of the FDS guideline, and identifying necessary revisions of the guideline.

4.2.1 Objectives of the Pilot Study Plan

1. To measure the clinical benefits of the proposed guideline, in terms of triage reliability

2. To test out the feasibility of integrating FDS into current triage practice

3. To identify difficulties encountered when performing FDS

4. To collect staff opinions and feedbacks, and to determine necessary revision of the

37

guideline

4.2.2 Time Frame

The pilot study will take place in the proposed AED setting in a public hospital of the

Hong Kong East cluster. As mentioned in chapter 3, the total attendance of the AED in 2012

is 146,000. Among those patients, approximately 10% were admitted with chest pain as their

chief compliant. By calculation, there are 40 potential eligible cases per day. The target

number of subject recruitment is 100. The target duration of the pilot study plan is within 1

month, but depends on when the target number of subjects is achieved.

4.2.3 Patient Selection

The inclusion and exclusion criteria of the subjects in the pilot study will be identical to

that of the proposed FDS guideline. All recruited subjects should be of aged 18 or above,

presenting to the AED with undifferentiated chest pain as their major complaint. Those with

chest pain of non-cardiac origins such as trauma or surgery, and those with obvious

ST-elevation in first ECG will be excluded.

4.2.4 Method

A well written guideline of FDS flow chart is available at triage station for reference. The

triage nurses will recruit eligible participants according to the guideline. A verbal consent

from the participants will be obtained. The triage nurses will perform triage, order ECG to the

participants, calculate FDS, and categorize participants according to the risk stratification and

38

other clinical data. All demographic data, vital signs, corresponding FDS score, triage

category, the provisional and final diagnosis of the study subjects and their final disposal will

be recorded in AED notes. For patients who are admitted to the hospital, their discharge

records, information on any cardiac procedures done and other patients outcomes can be

retrieved from hospital electronic patient record (ePR) and can be used for subsequent

analysis.

4.2.5 Evaluation of the Pilot Study

After implementation of the 1-month pilot study, the evaluation of the pilot study will be

performed. The aim of doing the evaluation is to determine whether the objectives of the pilot

study have been achieved. The communication team will review the data collected during the

pilot study. The team will review all the data of the recruited subjects by retrospectively

retrieve patient data from electronic patient record (ePR). The team will analyze data

regarding the recruitment of participants, FDS score calculated, assignment of triage category

based on FDS risk stratification, subsequent patient management, triage reliability, any

adverse outcomes of patient that possible attributed by the study or any other unforeseen

outcomes happened both for patient or for staff involved.

Furthermore, the team will arrange meetings to meet frontline nursing staff, so as to

gather feedbacks from them. Staff will be encouraged to express their feelings and concerns

about the pilot study, share their opinions, and point out the obstacles that they encountered

39

during the pilot study.

Based on the staff feedbacks, and the data analysis on the FDS effectiveness and patient

outcomes, the communication team will identify necessary revisions of the FDS guideline. A

written report of the pilot study will be handover to CREC as well as to AED administrative

staff for assessment. Formal launch of the FDS protocol will proceed after approvals have

been obtained from CREC & administrative staff.

4.3 EVALUATION PLAN

4.3.1 Identifying Outcomes

The implementation of FDS guideline is expected to increase the triage accuracy for

patients present to AED with undifferentiated chest pain. It is expected to have the

subsequent potential in improvement for the followings: (1) Early identification of chest pain

patient that result from Acute Coronary Syndrome (ACS) or Angina, (2) Improve triage

accuracy and sort out high risk patients, (3) Allow those high risk patients to achieve earlier

assessments & interventions, (4) Improve patient outcomes in term of morbidity and

mortality.

Patient Outcomes

The primary outcome variable for the study is the accuracy of triage category based on

FDS assessment. It is partly reflected by the provisional diagnosis in AED note or hospital

discharge diagnosis, which can be retrieved from ePR. High risk group patients (FDS score

40

5-6) are expected to have high chance of chest pain that are likely due to cardiac in origin, if

the provisional diagnosis from AED is ACS/Angina/Non-ST segment Elevation Myocardial

Infarction (NSTEMI), it show a positive correlation with urgent triage category and

intervention. In addition, the needs for those high risk group patients to receive thrombolytic

therapy, refer for urgent Percutaneous Coronary Intervention (PCI) or admit to Critical

Cardiac Unit (CCU) / Intensive Care Unit (ICU) also reflect the triage accuracy &

appropriateness of FDS guideline.

Comparison between the triage accuracy among patients of low risk group (FDS 0-2),

intermediate risk group (FDS 3-4) and high risk group (FDS 5-6) will also be done. The

accuracy can be reflected from the hospital discharge diagnosis or provisional diagnosis from

AED, as well as the need for urgent interventions like PCI or thrombolytic treatments. It

reflects the sensitivity and specificity for the risk stratification of FDS. For example if low

risk group patient is discharged from AED with provisional diagnosis of chest pain or

atypical chest pain and no immediate intervention is need, this reflects that the low risk group

is associated with low chance of adverse cardiac outcomes.

Healthcare Provider Outcomes

Healthcare provider outcomes can be reflected by the staff satisfaction regarding the

utilization of FDS guideline and their perceived skills, knowledge and confidence on

performing chest pain triage. As described in chapter one, many AED nurses in Hong Kong

41

are feeling uncertain in triage decision making, resulting in increased stress among them. As

FDS guideline is expected to aid chest pain triage decision making, the effectiveness of using

FDS guideline to increase staff morale, especially for triage nurses is one of our major

concerns.

System Outcomes

Inaccurate triage decisions can lead to inappropriate utilization of health resources and

adverse patient outcomes (Gerdtz & Bucknall, 1999). An incorrect triage category allocation

can lead to delay in treatment and death of a patient at worst. Such outcomes may have legal

consequences. FDS guideline is expected to minimize such adverse consequences.

Furthermore, FDS guideline is expected to identify high risk patients to achieve earlier

assessments and interventions, which in turn improve patients outcomes in terms of mortality

and morbidity, healthcare costs are expected to be reduced in long term.

4.3.2 When and How Often to take Measurement

The target subjects are patients present to AED with undifferentiated chest pain as chief

complaint. The recruitment of subjects will be done by triage nurses in all shifts. The triage

nurses calculate FDS, assign triage category according to FDS calculation and fill up the FDS

data entry form part 1 (see Appendix 6).

The data entry forms will be collected by working group. The outcome data will be

recorded by investigator in part 2 of the FDS data entry form (see Appendix 6). The

42

investigator will retrieve the corresponding AED diagnosis from Accident & Emergency

Information System (AEIS) and hospital discharge diagnosis from electronic patient record

(ePR) and record the data for subsequent analysis. The analysis will start after the target

number of subjects is achieved in the pilot study.

4.3.3 Data Analysis

The primary outcome variable for the study is the accuracy of triage based on FDS

assessment. It is reflected by the correlation of FDS risk grouping to hospital discharge

diagnosis or provisional diagnosis from AED e.g. ACS/Angina/Non-ST segment Elevation

Myocardial Infarction (NSTEMI). In addition, the correlation with the need for urgent

interventions like PCI or thrombolytic therapy also reflects the correlation agreement.

The triage reliability and accuracy of using FDS is our main evaluation objectives, which

can be analyzed by using Kappa (κ) statistics. Kappa (κ) statistic is used as a quantitative

measure of the magnitude of agreement between FDS risk grouping, triage categories and

final diagnosis of patient. The calculation is based on the difference between "observed"

agreement compare to "expected" agreement. The following guideline will be used for the

interpretation of κ statistic: κ of <0, less than chance agreement; κ of 0.01-0.20, slight

agreement; κ of 0.21 – 0.40, fair agreement; κ of 0.41-0.60, moderate agreement; κ of

0.61-0.80, substantial agreement; κ of 0.81-0.99, almost perfect agreement. The FDS is

considered successful if the agreement rate is of substantial or above, i.e. κ > 0.61.

43

The target subjects of the FDS protocol will be the consecutive eligible patients

presenting to the AED with undifferentiated chest pain as their chief compliant. Evaluation of

the FDS protocol will be performed after the implementation of the pilot study, subsequent

evaluations will be done annually after the formal launch of the protocol.

CONCLUSION

The objectives of the FDS guideline are to provide a practical and systematic approach to

perform chest pain triage; to stratify the risk in patients with undifferentiated chest pain; sort

out those high risk patients for early assessments & interventions and to improve triage

accuracy for patients presenting with undifferentiated chest pain in AED. When considering

the effectiveness of the guideline, triage reliability is assessed. The FDS guideline is

considered effective if the patient outcomes (diagnosis of ACS, Angina, Non STEMI and

need for urgent intervention like thrombolytic or PCI) show a statistically significant

correlation with high FDS scoring. The agreement rates between FDS scoring, triage category

with the final provisional diagnosis is also suggestive of the success of the FDS guideline as a

standard tool to enhance chest pain triage accuracy in AED.

44

References

American Heart Association. (2013). ACC/AHA guidelines for the management of patients

with ST-elevation myocardial infarction. American Heart Association: Dallas.

Chase, M., Robey, J. L., Zogby, K. E., Sease, K. L., Shofer, F. S., & Hollander, J. E. (2006).

Prospective validation of the thrombolysis in myocardial infraction risk score in the

emergency department chest pain population. Annuals of Emergency Medicine, 48(3),

252-259.

Chung, Y. M. (2005). An exploration of accident and emergency nurse experiences of triage

decision making in Hong Kong. Accident and Emergency Nursing, 13(4), 206-213.

Cullen, L., Greenslade, J., Hammett, C. J., Brown, A., Chew, D., Bilesky, J., Than, M.,

Lamanna, A. L., Kimderley, R., Chu, K., & Parsonage, W. A. (2013). Comparison of three

risk stratification rules for predicting patients with acute coronary syndrome presenting to

an Australian emergency department. Heart, Lung and Circulation, In Press, Corrected

Proof, 1-8.

Department of Health Hong Kong. (2010). Annual report 2009/2010. Retrieved August, 2014,

from http://www.dh.gov.hk/english/pub_rec/pub_rec_ar/pub_rec_arpis_0910_html.html

Gerdtz, M.F., & Bucknall, T.K. (1999). Why we do the things we do: applying clinical

decision-making framework to triage practice. Accident and Emergency Nursing, 7 (1),

50-57.

45

Goransson, K., Ehrenberg, A., Marklund, B., & Ehnfors, M. (2005). Accuracy and

concordance of nurses in emergency department triage. Scandinavian Journal of Caring

Sciences, 19(4), 432-438.

Ho, K. M., & Suen, K. P. (2013). Effectiveness of using the front door score to enhance the

chest pain triage accuracy of emergency nurse triage decisions. Journal of Cardiovascular

Nursing, 28(1), 1-10.

Hospital Authority (2011). Triage guideline (2011 edition). Hospital Authority: Hong Kong.

Hospital Authority (2012). HA Statistical Report 2011-2012. Retrieved August 2014, from

http://www.ha.org.hk/gallery/ha_publications.asp?Library_ID=15

Lee, B. S., Chang, A. M., Matsuura, A. C., Marcoon, S., & Hollander, M. D. (2011).

Comparison of cardiac risk scores in ED patients with potential acute coronary syndrome.

Critical Pathways in Cardiology, 10(2), 64-68.

Lyon, R., Morris, A. C., Caesar, D., Gray, S., & Gray, A. (2007). Chest pain to the emergency

department-to stratify risk with GRACE or TIMI? Resuscitation, 74, 90-93.

Morris, A. C., Caesar, D., Gray S. & Gray, A. (2006). TIMI risk score accurately risk

stratifies patients with undifferentiated chest pain presenting to an emergency department.

British Heart Journal, 92(9), 1333-1334.

Pollack, C. V., Sites, F. D., Shofer, F. S., Sease, K. L., & Hollander, J. E. (2006). Application

of the TIMI risk score for unstable angina and non-ST elevation acute coronary syndrome

46

to an unselected emergency department chest pain population. Academic Emergency

Medicine, 13, 13-18.

Ramsay, G., Podogrodzka, M., McClure, C., & Fox, K. A. A. (2007). Risk prediction in

patients presenting with suspected cardiac pain: the GRACE and TIMI risk scores versus

clinical evaluation. The Quarterly Journal of Medicine, 100(1), 11-18.

Scher, H. E. (1995). Chest pain: rapid assessment skills. Journal of Orthopedic Nursing,

14(3), 30-34.

Speake, D. (2003). Detecting high-risk patients with chest pain. Journal of Emergency Nurse,

11(5), 19-21.

World Health Organization (2008). The global burden of disease: 2004 update. Retrieved August,

2014 from http://www.who.int/healthinfo/global_burden_disease/2004_report_update/en/

47

Appendix 1

Search history

Search Date: 15 August 2013

Database

Search Terms

CINAHL Ovid

MEDLINE

PubMed

1. Chest pain 6,060 27,083 70,866

2. Acute coronary syndrome 3,206 12,753 16,503

3. Angina 6,831 59,407 64,448

4. 1 OR 2 OR 3 14,403 89,959 96,452

5. Risk stratification 2,529 13,086 17,848

6. Risk score 2,628 11,244 68,573

7. 5 OR 6 4,966 23,403 84,029

8. Emergency department 19,461 64,325 108,984

9. 4 AND 7AND 8 80 293 422

(4 AND 7 AND 8) NOT CT

NOT computer tomography

NOT computed tomography

NOT troponin

43 157 188

Topic screened and Abstract Read 2 7 9

Eliminated duplicate 1 1 6

48

Appendix 2

Table of Evidence Bibliographic

citation

Study type Participants &

characteristics

Assessment tool Comparison Outcome measures

Results/conclusions

Ho et al., (2013) Cross-

sectional

descriptive

study

n= 200

Patients attending

AED in HK of:

>18 year-old

Chinese

Chest pain as chief

complaint

Mean age 57.3

Front Door Score

(FDS)

(Simplified TIMI

risk score without

serum cardiac

marker component)

TIMI risk score

5-level triage

scale

To justify the appropriateness of risk

stratification

Gold standard: final physician diagnoses

Triage reliability using FDS, TIMI risk

score & 5-level triage scale were compared.

The agreement rates among the final

physician diagnoses, FDS, TIMI risk score,

and 5-level triage scale were computed

using k statistics.

Interpretation of k statistics:

k<0.41 poor to fair

k 0.41 to 0.60 moderate

k 0.61 to 0.80 substantial

k>0.80 almost perfect

Triage reliability

FDS 84%

TIMI risk score 84%

5-level triage scale 33.5%

k statistics of agreement rates

FDS k=0.696 (substantial)

TIMI risk score k=0.696 (substantial)

5-level triage scale k=0.063 (poor)

P< 0.001

The chest pain triage reliability of nurses

using FDS or TIMI risk score were more

credible than that of nurses using 5-level

triage scale.

Morris et al.,

(2006)

Cohort study n=980

Patients attending

AED in UK of:

>20 year-old

Chest pain as chief

complaint

Mean age 59

Front Door Score

(FDS)

(Simplified TIMI*

risk score without

serum cardiac

marker component)

TIMI risk score Outcome measures (within 30 days):

AMI

ACS

Angioplasty

Death

Comparison of the number of patients

experienced major cardiac events (refer to

the outcome measures) in each score value

group.

Diagnostic accuracy of FDS and TIMI risk

score were compared by calculation and

comparison of receiver operating

characteristic (ROC) curves and area under

the curves.

Area under the ROC curves:

FDS= 0.70

(95%CI 0.65-0.75)

TIMI risk score= 0.79

(95% CI 0.75-0.84)

P<0.01

TIMI risk score accurately risk stratifies

patients with undifferentiated chest pain.

FDS is not as sensitive or specific as TIMI

risk score, but can aid triage decision

making where cardiac marker is not

initially available.

49

Bibliographic

citation

Study type Participants &

characteristics

Assessment tool Comparison Outcome measures

Results/conclusions

Chase et al.,

(2006)

Prospective

observational

cohort study

n=1458

Patients attending

AED in US of:

>30year-old

Chest pain as

primary compliant

Mean age=53.2

TIMI risk score / Outcome measures (within 30 days)

AMI

Revascularization

All-cause mortality

Incidence:

TIMI 0, 1.7%

TIMI 1, 8.2%

TIMI 2, 8.6%

TIMI 3, 16.8%

TIMI 4, 24.6%

TIMI 5, 37.5&

TIMI 6, 33.3%

Highly significant relationship between

calculated TIMI risk score and 30-day

death, AMI and revascularization was

observed.

(χ², p<0.001)

Pollack et al.,

(2006)

Prospective

observational

cohort study

n=3929

Patients attending

AED in US of:

Chest pain

>24 year-old

TIMI risk score / Outcome measures (within 30 days):

AMI

Death

Revascularization

Incidence:

TIMI 0, 2.1%

TIMI 1, 5%

TIMI 2, 10.1%

TIMI 3, 19.5%

TIMI 4, 22.1%

TIMI 5, 39.2%

TIMI 6, 45%

TIMI 7, 100%

TIMI risk score correlated to the

likelihood of adverse outcomes within

30 days.

(chi-square, p<0.001)

50

Bibliographic

citation

Study type Participants &

characteristics

Assessment tool Comparison Outcome measures

Results/conclusions

Ramsay et al.,

(2007)

Prospective

observational

study

n=347

Patients attending

AED in UK of:

Chest pain

>18 year-old

GRACE risk score TIMI risk score

Clinical

parameters

Outcome measures (3 months):

All-cause mortality

Revascularization

ACS

Stroke

AED re-admission

Outcomes were plotted in receiver

operating characteristic (ROC) curves.

(sensitivity versus 1-specificity)

Diagnostic accuracy of GRACE and TIMI

risk score were compared by calculation

and comparison of ROC curves and area

under the curves.

Area under the ROC curves:

GRACE risk score= 0.82

TIMI risk score= 0.74

Clinical parameters= 0.55

Both GRACE and TIMI risk score can

predict adverse outcomes in unselected

patients presenting with chest pain in

AED.

The GRACE risk score is more superior

to TIMI risk score.

Lyon et al.,

(2007)

Descriptive

study

n=760

Patients attending

AED in UK of:

Chest pain

>20 year-old

Mean age=68

GRACE risk score TIMI risk score Outcome measures (30-day):

(30-day major cardiac outcome)

AMI

ACS

PCI

All-cause mortality

Diagnostic accuracy of GRACE and TIMI

risk score were compared by calculation

and comparison of ROC curves and area

under the curves.

Area under the ROC curves:

GRACE risk score= 0.8

(95% CI 0.75-0.85)

TIMI risk score= 0.79

(95% CI 0.74-0.85)

P< 0.01

GRACE and TIMI risk score have

similar sensitivity and specificity in

predicting outcomes in patients

presenting to AED with undifferentiated

chest pain.

51

Bibliographic

citation

Study type Participants &

characteristics

Assessment tool Comparison Outcome measures

Results/conclusions

Cullen et al.,

(2013)

Prospective

observational

study

n=948

Patients attending

AED in Australia

of:

At least 5-min of

chest pain

HFA/CS-ANZ

guideline

GRACE score

TIMI risk score

Outcome measures (30-day):

All-cause death

AMI

Unstable angina

Diagnostic accuracy of HFA/CS-ANZ

guideline, GRACE and TIMI risk score

were compared by calculation and

comparison of ROC curves and area under

the curves.

Area under the ROC curves:

HFA/CS-ANZ= 0.75

(95% CI 0.72-0.81)

GRACE risk score= 0.83

(95% CI 0.79-0.87)

TIMI risk score= 0.79

(95% CI 0.74-0.83)

P<0.01

Three tools for risk stratification of ACS

were similar in their performance.

Complexity of GRACE risk score limits

its utility.

Lee et al., (2011)

Prospective

cohort study

n=4743

Patients attending

AED in US of:

Chest pain

>30 year-old

Mean age= 52.5

TIMI risk score GRACE score

PURSUIT score

Outcome measures (30 days):

Death

AMI

Revascularization

Predictive accuracy of TIMI risk score,

GRACE risk score and PURSUIT risk

score were compared by calculation and

comparison of ROC curves and area under

the curves.

Area under the ROC curves:

TIMI risk score= 0.757

(95% CI 0.728-0.785)

GRACE risk score= 0.728

(95% CI 0.701-0.755)

PURSUIT risk score= 0.691

(95% CI 0.662-0.720)

TIMI risk score had the best ability to

predict 30-day cardiovascular events.

Abbreviations: AED, accidents and emergency departments; FDS, Front Door Score; TIMI, Thrombolysis in Myocardial Infarction; AMI, acute myocardial infarction; ACS,

acute coronary syndrome; ROC, receiver operating characteristic; GRACE, the Global Registry of Acute Coronary Events; HFA/CS-ANZ, the Heart Foundation of Australia

and Cardiac Society of Australia and New Zealand; PURSUIT, Platelet glycoprotein IIb/IIIa in Unstable angina: Receptor Suppression Using Integrilin; PCI, Percutaneous

coronary intervention; CABG, coronary artery bypass graft

52

Appendix 3

SIGN Appraisal Checklists

Ho, K. M., & Suen, K. P. (2013). Effectiveness of using the front door score to enhance the

chest pain triage accuracy of emergency nurse triage decisions. Journal of Cardiovascular

Nursing, 28(1), 1-10.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Can’t say

4. The included patients and settings match the key question. Yes

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Can’t say

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Can’t say

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Can’t say

13. All patients recruited into the study are included in the analysis. Yes

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? Acceptable +

15. What is your assessment of the applicability of this study to our target population? Directly applicable

16. Notes. Summarize the authors conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

The chest pain triage reliability of the nurses using TIMI risk score or Front Door Score was found to be more credible than those using 5-level

triage scale, represented by the agreement rates against the final physician diagnosis found in this study.

Some bias in the study design was seen, e.g. only Chinese patients are included, blinding method was not mentioned for the investigators and

the attending physician, manipulation of calculation in score could occur, discrepancies among different physicians’ judgments could exist.

Calculation of Front Door Score together with current 5-level triage scale is feasible and may enhance chest pain triage accuracy.

Level of evidence +

53

Morris, A. C., Caesar, D., Gray S. & Gray, A. (2006). TIMI risk score accurately risk

stratifies patients with undifferentiated chest pain presenting to an emergency department.

British Heart Journal, 92(9), 1333-1334.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Yes

4. The included patients and settings match the key question. Yes

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Can’t say

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Can’t say

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Yes

13. All patients recruited into the study are included in the analysis. Yes

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? Acceptable +

15. What is your assessment of the applicability of this study to our target population? Directly applicable

16. Notes. Summarize the authors conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

TIMI risk score accurately predicts the 30-day major cardiac events rate among undifferentiated chest pain population in AED. Front Door

Score is not as sensitive or specific as TIMI risk score, but can aid triage decision making where cardiac marker is not initially available.

Strengths of the study: Data completeness 95%, follow up of patients arranged, outcome measures clearly defined.

Limitations: Only statistics of composite endpoints (combined all outcome measures) was presented, methods to minimize bias of interpreters/

researchers was not mentioned, manipulation of calculation in score could occur.

The validity of Front Door Score was examined, and was found to be useful to aid triage decision making, as well as subsequent patient

management.

Level of evidence +

54

Chase, M., Robey, J. L., Zogby, K. E., Sease, K. L., Shofer, F. S., & Hollander, J. E. (2006).

Prospective validation of the thrombolysis in myocardial infraction risk score in the

emergency department chest pain population. Annuals of Emergency Medicine, 48(3),

252-259.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Yes

4. The included patients and settings match the key question. Yes

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Yes

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Yes

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Yes

13. All patients recruited into the study are included in the analysis. Yes

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? High quality ++

15. What is your assessment of the applicability of this study to our target population? Directly applicable

16. Notes. Summarize the authors conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

Highly significant relationship between calculated TIMI risk score and 30-day death, AMI and revascularization was observed.

Strength: Large sample size, different personnel responsible for calculation of score and follow up outcomes, methods to minimize bias

mentioned, samples with incomplete data also being analyzed (intention to treat), outcome measures clearly defined.

Level of evidence ++

55

Pollack, C. V., Sites, F. D., Shofer, F. S., Sease, K. L., & Hollander, J. E. (2006). Application

of the TIMI risk score for unstable angina and non-ST elevation acute coronary syndrome to

an unselected emergency department chest pain population. Academic Emergency Medicine,

13, 13-18.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Yes

4. The included patients and settings match the key question. Yes

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Can’t say

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Can’t say

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Yes

13. All patients recruited into the study are included in the analysis. Yes

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? Acceptable +

15. What is your assessment of the applicability of this study to our target population? Directly applicable

16. Notes. Summarize the author’s conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

TIMI risk score correlates to the likelihood of adverse outcomes in 30 days, it can be used as a tool to risk stratify patient attending AED for

chest pain.

Strength: Large sample size, samples with incomplete data also being analyzed (intention to treat).

Limitations: Methods to minimize bias of interpreters/ researchers was not mentioned, manipulation of calculation in score could occur.

Level of evidence +

56

Ramsay, G., Podogrodzka, M., McClure, C., & Fox, K. A. A. (2007). Risk prediction in

patients presenting with suspected cardiac pain: the GRACE and TIMI risk scores versus

clinical evaluation. The Quarterly Journal of Medicine, 100(1), 11-18.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Yes

4. The included patients and settings match the key question. Can’t say

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Can’t say

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Can’t say

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Yes

13. All patients recruited into the study are included in the analysis. Yes

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? Acceptable +

15. What is your assessment of the applicability of this study to our target population? Some indirectness

16. Notes. Summarize the author’s conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

Both GRACE and TIMI risk scores were able to predict 30-day adverse outcome events.

GRACE score was found to perform better than TIMI risk score in risk stratification. However it is more complicated in calculating, online

calculator is needed to do the calculation.

Limitation: The target population in this study is those with suspected cardiac pain in AED and cardiac ward, methods to minimize bias of

interpreters/ researchers was not mentioned, manipulation of calculation in score could occur.

Level of evidence +

57

Lyon, R., Morris, A. C., Caesar, D., Gray, S., & Gray, A. (2007). Chest pain to the emergency

department-to stratify risk with GRACE or TIMI? Resuscitation, 74, 90-93.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Yes

4. The included patients and settings match the key question. Yes

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Can’t say

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Can’t say

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Yes

13. All patients recruited into the study are included in the analysis. No

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? Acceptable +

15. What is your assessment of the applicability of this study to our target population? Directly applicable

16. Notes. Summarize the author’s conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

GRACE and TIMI risk score were found to have similar sensitivity and specificity in predicting 30-day major cardiac event. Both can risk

stratify patients attending AED for chest pain.

Limitation: some confounding factors can be seen in the study design, e.g. GRACE and TIMI risk score were calculated retrospectively, 24%

of recruited eligible sample were excluded due to incompleteness of data to calculate the score, methods to minimize bias of interpreters/

researchers was not mentioned, manipulation of calculation in score could occur.

Strength: GRACE and TIMI risk score were compared by each score group respectively

Level of evidence -

58

Cullen, L., Greenslade, J., Hammett, C. J., Brown, A., Chew, D., Bilesky, J., Than, M.,

Lamanna, A. L., Kimderley, R., Chu, K., & Parsonage, W. A. (2013). Comparison of three

risk stratification rules for predicting patients with acute coronary syndrome presenting to an

Australian emergency department. Heart, Lung and Circulation, In Press, Corrected Proof,

1-8.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Yes

4. The included patients and settings match the key question. Yes

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Can’t say

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Can’t say

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Yes

13. All patients recruited into the study are included in the analysis. Yes

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? High quality++

15. What is your assessment of the applicability of this study to our target population? Directly applicable

16. Notes. Summarize the author’s conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

HFA/CS-ANZ, GRACE and TIMI risk score were similar in performance in predicting the risk of 30-day ACS, death, AMI and unstable

angina in patients attending AED for chest pain.

Strength: Very detailed descriptions on the definition of outcome measures were reported, objectives data obtained to measure outcomes,

detailed diagnostic accuracy statistics (on sensitivity, specificity, negative predictive value, positive predictive value) for the three scores were

presented.

GRACE risk score has higher discriminatory power to differentiate ACS and non-ACS patients, but it is more complex to calculate.

Level of evidence ++

59

Lee, B. S., Chang, A. M., Matsuura, A. C., Marcoon, S., & Hollander, M. D. (2011).

Comparison of cardiac risk scores in ED patients with potential acute coronary syndrome.

Critical Pathways in Cardiology, 10(2), 64-68.

PATIENT SELECTION

1. A consecutive sequence or random selection of patients is enrolled. Yes

2. Case – control methods are not used. Yes

3. Inappropriate exclusions are avoided. Yes

4. The included patients and settings match the key question. Yes

INDEX TEST

5. The index test results interpreted without knowledge of the results of the reference standard. Can’t say

6. If a threshold is used, it is pre-specified. Not applicable

7. The index test, its conduct, and its interpretation is similar to that used in practice with the target

population of the guideline.

Yes

REFERENCE STANDARD

8. The reference standard is likely to correctly identify the target condition. Yes

9. Reference standard results are interpreted without knowledge of the results of the index test. Can’t say

10. The target condition as defined by the reference standard matches that found in the target population of

the guideline.

Yes

FLOW AND TIMING

11. There is an appropriate interval between the index test and reference standard. Yes

12. All patients receive the same reference standard. Yes

13. All patients recruited into the study are included in the analysis. Yes

OVERALL ASSESSMENT OF THE STUDY

14. How well was the study done to minimize bias? Acceptable+

15. What is your assessment of the applicability of this study to our target population? Directly applicable

16. Notes. Summarize the author’s conclusions. Add any comments on your own assessment of the study, and the extent to which it answers

your question.

TIMI risk score best predict 30-days adverse cardiovascular events in unselected patients attending AED for chest pain

Strength: Very large sample size, not only composite outcomes being measured, statistical results for individual outcome measures for each

score were reported.

Level of evidence ++

60

Appendix 4

Scottish Intercollegiate Guidelines Network SIGN

SIGN GRADING SYSTEM 1999 – 2012

LEVELS OF EVIDENCE

1++ High quality meta-analyses, systematic reviews of RCTs, or RCTs with a very low risk of bias

1+ Well-conducted meta-analyses, systematic reviews, or RCTs with a low risk of bias

1- Meta-analyses, systematic reviews, or RCTs with a high risk of bias

2++ High quality systematic reviews of case control or cohort or studies

High quality case control or cohort studies with a very low risk of confounding or bias and a high

probability that the relationship is causal

2+ Well-conducted case control or cohort studies with a low risk of confounding or bias and a moderate

probability that the relationship is causal

2- Case control or cohort studies with a high risk of confounding or bias and a significant risk that the

relationship is not causal

3 Non-analytic studies, e.g. case reports, case series

4 Expert opinion

GRADES OF RECOMMENDATIONS

A At least one meta-analysis, systematic review, or RCT rated as 1++, and directly applicable to the target

population; or

A body of evidence consisting principally of studies rated as 1+, directly applicable to the target

population, and demonstrating overall consistency of results

B A body of evidence including studies rated as 2++, directly applicable to the target population, and

demonstrating overall consistency of results; or

Extrapolated evidence from studies rated as 1++ or 1+

C A body of evidence including studies rated as 2+, directly applicable to the target population and

demonstrating overall consistency of results; or

Extrapolated evidence from studies rated as 2++

D Evidence level 3 or 4; or

Extrapolated evidence from studies rated as 2+

61

Appendix 5

Costs of Implementation

Durations Man-hour

Staff

15 APN/NO 1-hr training session 15

53 RN 1-hr training session 53

Total man-hour 68

Cost of Implementation (Material costs)

Items Costs & Amount Total Costs

Printing costs $0.1 x 14,600 $1,460

Venue cost --- ---

Total Cost = $1,460/year

62

Appendix 6

Accident & Emergency Department

A standardized Protocol to Increase Triage Accuracy for Patient Presenting with Cardiac

Symptoms in AED

Data Entry Form (Part 1)

Please complete the checklist if the patient present to AED with chief complaint as chest

pain

- Inclusion criteria:

Age of 18 year or above, stable vital sign at initial triage

- Exclusion criteria:

Aged below 18

Mentally incapable

Obvious non-cardiac origin account for the chest pain

Critically ill or hemodynamically unstable at presentation

Obvious ST-segment elevation Myocardial Infarct (STEMI) in triage ECG

Vital Sign at presentation:

Blood Pressure____________, Pulse___________, Respiratory Rate____________,

SpO2 (Room Air / ____L/% O2) __________, Temperature _________

Door-to-ECG time ____________

Conversion & Incorporate FDS risk stratification category group into triage category

Calculate total FDS Score 0-2 3-4 5-6

Stratify risk category base on FDS

score for having adverse cardiac

outcome

Low risk Intermediate

risk

High risk

Assigned triaged category

correspond to FDS stratify risk

Category 4

Semi urgent

Category 3

Urgent

Category 2 or 1

Critical or

Emergency

63

Front Door Score (FDS)

Prognostic Variables Yes / No Score

1) Age ≥ 65 years No 0

Yes 1

2) Previous coronary stenosis ≥ 50% (select at least 1 of

the below to gain 1 point)

Previous MI, Previous PTCA, Previous CAGB

No 0

Yes 1

3) At least 3 risk factors for CHD (select at least 3 factors to

gain 1 point)

Hypertension, Diabetes, Hypercholesterolemia, Family

history of CHD, History of tobacco use

No 0

Yes 1

4) Use of aspirin in the previous 7 days No 0

Yes 1

5) At lease 2 angina episodes in the last 24 hour No 0

Yes 1

6) ST-segment depression or elevation ≥ 0.5mm No 0

Yes 1

FDS Total Score

Assigned triaged category correspond to FDS stratify risk

Data Entry form (Part 2) Patient is: Admitted to Hospital / Discharged by AED *

Provisional Diagnosis in AED

Acute Myocardial Infarct (AMI)

Acute Coronary Syndrome (ACS)

Unstable Angina

Suspected Stable Cardiac Chest Pain

Suspected Non-Cardiac Chest Pain

Others (please specify):

* Please delete inappropriate

For patient that is admitted to Hospital:

Yes (please specify) No

Any Admission to ICU / CCU?

Cardiac Intervention e.g. thrombolytic, PCI?

Any adverse cardiac outcome document?

Hospital Discharge Diagnosis

64