a b s t r a c t

Objective: The HAS-Choice pathway utilizes the HEART score, an accelerated Diagnostic protocol (ADP), and shared decision-making using a visual aid in the evaluation of chest pain patients. We seek to determine if our intervention can improve resource utilization in a community emergency department (ED) setting while main- taining safe patient care.

Methods: This was a single-center prospective cohort study with historical that included ED patients >=21 years old presenting with a primary complaint of chest pain in two time periods. The primary outcome was patient dispo- sition. Secondary outcomes focused on 30-day ED bounce back and Major adverse cardiac events . We used multivariate logistic regression to estimate the odds ratio (OR) and its 95% confidence interval (CI).

Results: In the pre-implementation period, the unadjusted disposition to inpatient, observation and discharge was 6.5%, 49.1% and 44.4%, respectively, whereas in the post period, the disposition was 4.8%, 41.5% and 53.7%, respectively (chi-square p b 0.001). The adjusted odds of a patient being discharged was 40% higher (OR = 1.40; 95% CI, 1.30, 1.51; p b 0.001) in the post-implementation period. The adjusted odds of patient admission was 30% lower (OR = 0.70; 95% CI, 0.60, 0.82; p b 0.001) in the post-implementation period. The odds of 30- day ED bounce back did not statistically differ between the two periods. MACE rates were b1% in both periods, with a significant decrease in mortality in the post-implementation period.

Conclusion: Our study suggests that implementation of a shared decision-making tool that integrates an ADP and the HEART score can safely decrease hospital admissions without an increase in MACE.

(C) 2018

Introduction

Chest pain is one of the most common chief complaints among pa- tients presenting to the emergency department (ED), accounting for nearly 8 million visits in the United States each year [1]. Many of these patients undergo significant cardiac workups, yet b10% of these patients are ultimately diagnosed with acute coronary syndrome [2-5]. Emer- gency physicians must accurately identify patients who are either experiencing acute coronary syndrome (ACS) or at risk for ACS in the near future. Previously, AHA/ACC guidelines recommended that pa- tients undergo cardiac Stress testing within 72 h of ED presentation [6]. The high morbidity and mortality as well as risk of litigation

* Corresponding author at: 5301 East Huron River Drive, P.O. Box 995, Ann Arbor, MI 48106, USA.

E-mail addresses: ggafnipappas@epmg.com (G. Gafni-Pappas), sdemeester@epmg.com (S.D. DeMeester), gantia@med.umich.edu (A. Ganti), adni@umich.edu (A.M. Nicholson), jeremy@methodsconsultants.com (J. Albright), juan.wu@stjoeshealth.org (J. Wu).

stemming from missed ACS in the past lead to increased hospital and re- source utilization for chest pain patients [5].

Many institutions utilize accelerated diagnostic protocols to assess for ACS, which include EKGs and serial troponins. The decision regard- ing which patients were appropriate for discharge or admission was complicated because of lack of evidence. Thus, admission became the preferred disposition so that major adverse cardiac events were not missed. However, over the past decade, risk stratification tools such as the HEART score have shown that emergency physicians can safely risk stratify chest pain patients and decrease resource utilization [7- 10]. In fact, the most recent 2015 AHA/ACC guidelines support the use of risk stratification tools to identify Low-risk chest pain patients that may be suitable for outPatient evaluation [11]. The HEART score was developed specifically to risk stratify undifferentiated emergency de- partment chest pain patients, and has been externally validated in mul- tiple follow up studies [7,12-17]. Points are assigned for history, ECG, age, risk factors, and troponin results. The resulting score stratifies pa- tients into low, moderate, or high-risk groups. Patients with a HEART score of 0-3 have a b2% risk of a major adverse cardiac event

https://doi.org/10.1016/j.ajem.2018.02.005

0735-6757/(C) 2018

at 30 days. When combined with serial negative troponins, patients have been shown to have b1% risk for 30 day MACE [18]. Utilizing this HEART Pathway on chest pain patients, Mahler et al. found an increased early discharge rate of 21.3% without an increase in MACE [7]. Further- more, shared decision-making employing a visual aid for low-risk chest pain patients has been shown to improve patient satisfaction and un- derstanding [19,20].

Though academic centers have shown that using the HEART score can safely reduce admissions and increase satisfaction using a shared decision-making visual aid, diffusion into community emergency de- partments may be slow. Potential barriers to implementation could include lack of support for operational changes or for physician educa- tion, access to follow-up, or perceived inadequate time to engage in shared decision-making with patients during a busy shift. In addition, published randomized clinical trials (RCT) are usually conducted with a number of exclusion criteria, which can limit generalizability of the findings. In this study, the HAS-Choice pathway utilized the HEART Score, an accelerated diagnostic protocol (ADP), and shared decision- making using a visual aid in the evaluation of chest pain patients pre- senting to a community ED. By using the shared decision-making tool, incorporating the ADP and HEART score into the discussion of risk with the patient was more manageable for the clinician, creating simple protocol implementation. The main goal of our study is not to demon- strate its efficacy but to evaluate the effectiveness of implementation in a diverse patient population. We hypothesize that the pathway can improve patient care and resource utilization when implemented in a community ED setting. To our knowledge, this is the first study to com- bine the use of the HEART pathway with a shared decision-making tool.

Methods

Study design and setting

This is a prospective observational cohort study with historical con- trols conducted at Saint Joseph Mercy Hospital in Ann Arbor, MI. Saint Joseph Mercy Hospital is a 537 bed community hospital that sees ap- proximately 85,000 ED patients annually and is staffed by 40 attending physicians and 10 advanced practice providers. Certain areas of the ED also incorporate rotating EM residents from the University of Michi- gan/Saint Joseph Mercy Hospital emergency medicine residency. The Emergency Observation Center (EOC) is a closed 12-bed unit, managed by ED providers. The study was approved by the Institutional Review Board at Saint Joseph Mercy Health System.

Selection of participants

The HAS Choice intervention was implemented on January 1, 2016. We used an “intention-to-treat” approach to identify the study popula- tion in the pre and post-implementation period. Patients in the post- implementation period (prospective observational cohort) included all ED patients >=21 years old presenting with a primary complaint of chest pain concerning for acute coronary syndrome who underwent Troponin testing during calendar year 2016. Exclusion criteria included known pregnancy and age b21 years old. ICD-10 Diagnosis codes (see supplemental file for a comprehensive list of ICD-10 codes) were used to identify chest pain patients from hospital electronic medical records. Patients in the pre-implementation period (historical control) came from the entire year of 2014 with the same inclusion and exclusion criteria as 2016 except that ICD-9 diagnosis codes were applied due to the switch from ICD-9 to ICD-10 on October 1, 2015 (see supplemental file for a comprehensive list of ICD-9 codes).

HAS-Choice intervention

Before the implementation of the HAS-Choice intervention, a “Rapid Rule-Out” ADP was the standard of care in the evaluation of chest pain

patients at our institution. The ADP utilizes serial ECGs, two-hour serial troponins and a delta troponin (Seimens Centra ultra 3rd generation cTNi). If the delta troponin increases N0.02 ng/mL, the patient receives a third troponin at 6 h. Patients that ‘rule-in’ for NSTEMI are admitted to inpatient status, whereas other patients are either admitted to the EOC under Observation status for provocative testing or a cardiology consult, or discharged home.

The HAS-Choice pathway adds a validated clinical decision rule with a shared decision-making tool utilizing the HEART score (Figs. 1, 2, and 3) to the existing ADP. Once the initial ADP is complete, a color paper form of the shared decision-making tool is used to discuss risk with the patient and determine disposition. The form is signed by the physi- cian and patient to confirm that it is completed. All attending emer- gency physicians, residents, and advanced practice providers were trained regarding the HEART score and shared decision-making visual aid through an online module and multiple communications through email and presentations at department meetings.

Outcomes

The primary outcome of interest was patient disposition before and after implementation of the HAS-Choice pathway. The three disposi- tions were discharge, admit to observation (OBS), and admit to inpa- tient (IP). Discharge was used to describe initial disposition from the ED, not discharge from OBS or IP. The secondary outcomes were 30- day rates of ED bounce backs, MACE, and death before and after imple- mentation. ED bounce backs and MACE were identified from hospital data whereas deaths were determined from hospital data, Social Secu- rity and Michigan Death Indices. Our definition of MACE included non-ST elevation myocardial infarction (NSTEMI), ST-elevation myocar- dial infarction (STEMI), Percutaneous Transluminal Coronary Interven- tion/Percutaneous Coronary Intervention (PTCI/PCI), cardiac arrest, or death.

Statistical analysis

Demographic and clinical characteristics of patients were summa- rized by percentage (frequency) and mean +- SD, where appropriate. Chi-square tests (or Fisher’s Exact tests) and two-sample t-tests were used to compare the characteristics of patients presenting to the ED be- fore versus after implementation of the HAS-Choice pathway.

We used multivariate logistic regression to estimate the odds ratio (OR) with 95% confidence intervals (CI’s) for the association between intervention and disposition outcomes. Each disposition outcome [dis- charge, inpatient (IP), or observation (OBS)] was coded as a binary event. Specifically, the logistic regression model used patient disposi- tion as the dependent variable, the time period (pre/post) as the explan- atory variable, and adjusted for risk factors whose distribution varied between the pre- and post-period including age, BMI, race, Insurance type, and number of Cardiovascular risk factors. The same statistical method is applied for the secondary outcomes including ED bounce backs and adverse events.

We also analyzed a subgroup of patients in the post period from the HAS-Choice project database who had HEART scores documented in the chart (n = 2302). The relationship between a patient’s HEART score risk stratification and patient disposition were examined using grouped bar charts and Chi-square tests. All statistical tests were two-sided and con- ducted at a significance level of 0.05. All analyses were performed using SAS (version 9.4) and R (version 3.3.3).

Results

Characteristics of study subjects

The total sample size during the study period was 13,445 patients, 7657 in the pre-intervention cohort and 5788 in the post-intervention

Fig. 1. Shared decision-making visual aid for Low risk chest pain patients utilizing the HEART score.

Fig. 2. Shared decision-making visual aid for moderate risk chest pain patients utilizing the HEART score.

Fig. 3. Shared decision-making visual aid for high risk chest pain patients utilizing the HEART score.

cohort. Demographic and clinical characteristics are presented in Table 1. Significant differences were observed in age, BMI, race, insur- ance type, and history of cardiovascular risk factors (p-value b 0.001).

Main results

Overall, in the pre-implementation period, the unadjusted disposi- tion to IP, OBS and discharge was 6.5%, 49.1% and 44.4%, respectively, whereas in the post-period, the disposition was 4.8%, 41.5% and 53.7%,

Table 1

Comparison of demographic and clinical characteristics between the pre and post imple- mentation period.

respectively (chi-square p b 0.001) (Table 2). When we visualized a monthly rate for each disposition outcome, we also observed a general raw trend towards a discharge disposition (Fig. 4).

When further adjusted for patient characteristics (Table 2), com- pared to the pre-implementation period, the odds of a patient being discharged was 40% higher (adjusted OR = 1.40; 95% CI, 1.30, 1.51; p b 0.001) in the post-implementation period. Furthermore, the odds of patient admission to IP and OBS was 20-30% lower in the post implementation period.

In regards to adverse events (Table 3), there was an increase in the odds of 30-day ED bounce back in the post-implementation period but this was not statistically significant after multivariate adjustment. How- ever, the odds of 30-day mortality significantly decreased in the post pe-

riod (adjusted OR = 0.47; 95% CI, 0.25, 0.88; p = 0.02).

Variable Pre (n = 7657) Post

(n = 5788)

Categorical N (%) N (%)

Gender (male) Insurance type – non-medicare Race (non-white) Married	41.7% (3193) 54.9% (4209) 25.4% (1945) 49.7% (3806)	42.6% (2463) 62.4% (3614) 29.1% (1684) 49.3% (2854)	0.33 b0.001 b0.001 0.65	of ED bounce back in the post period. However, the odds of death was 65% (adjusted OR = 0.35; 95% CI, 0.14, 0.88; p = 0.03) lower in the post period. When stratifying by age group, patients N65 years old are
History of diabetes	14.5% (1113)	18.3% (1060)	b0.001	more likely to experience an ED bounce back. Furthermore, the risk of
History of hyperlipidemia	19.5% (1494)	24.5% (1417)	b0.001	ED bounce back was higher in the post period. Despite the increased
History of statins History of Hypertension history of acute MI	9.7% (744) 31.9% (2444) 3.1% (240)	15.7% (911) 39.7% (2300) 4.8% (281)	b0.001 b0.001 b0.001	risk of ED bounce back, the risk of death decreased in the post period. The risk of death or all adverse events were lower in all age groups in
History of stroke	2.0% (151)	2.8% (160)	0.003	the post period.
History of peripheral artery	2.7% (207)	3.9% (228)	b0.001	Fig. 5 focuses on a subgroup of patients who had HEART scores doc-
disease Continuous Age	Mean (Std. Dev.) 58.9 (16.7)	Mean (Std. Dev.) 55.3 (16.3)	t-Test b0.001	umented in the post-implementation period. Those who were missing the score had a higher risk and comorbidities than those who had the

p

Value

Table 4 focuses on those patients who were discharged directly from the ED in the pre- and post-period. Among all patients, there was a 14% increase (adjusted OR = 1.14; 95% CI, 0.99, 1.31; p = 0.07) in the odds

BMI	30.6 (7.3)	31.1 (7.4)	b0.001

For categorical variables, p values were calculated from chi-square or Fisher’s exact test, whichever appropriate. For continuous variables, p values were calculated from two inde- pendent sample t-test.

score documented in the ED (data not shown). Nevertheless, there was a statistically significant difference in patient disposition between the different HEART score risk groups (Fisher’s exact p b 0.001). Specif- ically, the proportion of discharge patients increased from high to low

Table 2

Disposition outcome comparison between pre and post period.

	Unadjusted				Adjusted
Disposition	Pre, % (N)	Post, % (N)	OR (95% CI) (REF = pre)	p Value	ORa (95% CI)	p Value
Discharge	44.4% (3401)	53.7% (3109)	1.45 (1.36, 1.56)	b0.001	1.40 (1.30, 1.51)	b0.001
OBS	49.1% (3760)	41.5% (2402)	0.74 (0.69, 0.79)	b0.001	0.78 (0.73, 0.84)	b0.001
Inpatient	6.5% (496)	4.8% (277)	0.73 (0.62, 0.84)	b0.001	0.70 (0.60, 0.82)	b0.001

Abbreviation: OR, odds ratio.

^a Multivariate logistic regression adjusted for age (b45 yrs, 54-64 yrs, >=65), BMI (b18.5, 18.5-24.9, 25.0-29.9, >=30.0 kg/m², and a missing indicator), insurance type (medicare vs non- medicare), race (white vs non-white), number of cardiovascular risk factors (0, 1-2, >=2).

HEART score groups whereas the admission to OBS exhibited the oppo- site pattern.

Discussion

This prospective cohort study with historical controls suggests that application of the HAS-Choice pathway incorporating an accelerated diagnostic protocol, HEART score risk stratification, and shared decision-making via a visual aid safely decreased admissions for pa- tients presenting to the emergency department with a primary com- plaint of chest pain at a community hospital. To our knowledge, this information has never before been reported in a community setting.

We chose an observational cohort study with historical controls to show the effectiveness of implementation within a community setting. ICD-9 codes were used in the 2014 historical control while ICD-10 codes were used for the observational cohort in 2016. We made a priori deci- sion not to use the year of 2015 as the historical control because we piloted the HAS-Choice intervention during this time at our institution. Furthermore, due to the increasing awareness and popularity of the HEART score in 2015, many of our providers had begun sporadically using this information for risk stratification prior to the full implemen- tation of the pathway. Thus, the patient population in 2014 would be most indicative of an uncontaminated, pre-intervention group.

With the escalating and unsustainable costs of healthcare, delivering care efficiently and cost-effectively is very important [21,22]. Using evidence-based protocols along with shared decision-making has become integral in assisting disposition practice [23]. Chest pain in par- ticular is a very common complaint in emergency departments. Prior studies demonstrate that the HEART score can identify chest pain pa- tients who can safely be discharged home, maintaining an acceptable low Adverse event rate [7,12,24]. Our study is unique in that the HAS- Choice intervention incorporates a shared decision-making tool to an al- ready established pathway that includes an ADP and the HEART score.

While the ADP and HEART score were important parts of the interven- tion, it was the integration of the shared decision-making tool that allowed ease of implementation at our community site.

Our ADP had already been in place in our hospital system for many years. Previously, physician gestalt was used to determine whether pa- tients were acceptable for Discharge home. By adding the HEART score and visual aid into clinical practice, we were able to show that there was a significant trend towards decreased admissions with a 40% in- creased odds of being discharged than admitted from pre- to post- implementation. Furthermore, the rate of adverse events was b1%, sim- ilar to previous studies, suggesting that implementation of the pathway is safe. The HAS-Choice pathway results are similar to other studies in that we have shown decreased utilization without increased MACE rates [7,12]. Though we did not study patient involvement and knowl- edge related to shared decision-making, our study was similar to Hess et al. in showing decreased admissions [19].

Despite the HEART score having been shown to reduce healthcare utilization [7], many emergency departments may not have developed effective methods to implement the risk stratification tool. Our protocol is feasible to implement at both academic and community emergency departments. It involves two components in addition to the standard accelerated Diagnostic pathway in an ED. These include 1) educating cli- nicians on use of the HEART score for the disposition of chest pain pa- tients, and 2) shared decision-making using a paper visual aid to discuss patient risk. It has been sustainable and well accepted by emer- gency providers, cardiologists, and primary care physicians at our institution.

Other studies have shown significant cost savings using the HEART score [25]. The impact on healthcare utilization at our institution shows a significant decrease in cost of care for chest pain patients. The average cost savings per person based on hospital charges is estimated to be around $1541.32. This is calculated based on the adjusted distribu- tion of disposition in the pre- and post-period. For our institution with

Fig. 4. The monthly rate of disposition to each route straddling pre and post period.

Table 3

Odds ratio for 30-day ED bounce back and adverse events comparing post to pre implementation period.

Variable	Pre (%, n)	Post (%, n)	Unadjusted p value	Adjusted ORsa (95% CI)	p Value
ED bounce back	13.63% (1044)	15.26% (883)	0.008	1.02 (0.92, 1.12)	0.76
Adverse events within 30 days of index ED visit NSTEMI or STEMI	0.05% (4)	0.12% (7)	0.22	NA	NA
PTCI/PCI	0.04% (3)	0.10% (6)	0.19	NA	NA
Cardiac arrest	0.04% (3)	0% (0)	0.26	NA	NA
Death	0.56% (43)	0.22% (13)	0.003	0.47 (0.25, 0.88)	0.02
Any adverse event	0.64% (49)	0.38% (22)	0.04	0.65 (0.39, 1.09)	0.10

NA indicates small sample size does not permit multivariate adjustment.

^a Logistic regressions were adjusted for the same variables as in the Table 2.

approximately 2500 patients presenting with chest pain, the savings is estimated to be $3,853,300 per year. Furthermore, by shifting the dispo- sition in favor of discharge compared to admit as well as observation compared to inpatient, length of stay can be decreased [7]. The impact is profound when generalized to the entire country.

Finally, the shared decision-making visual aid was perhaps the most important part of our protocol because it integrated the ADP and HEART score to allow straightforward and sustainable implementation. Empowering clinicians to discuss risk with their patients and patients to make an informed decision based on their risk can have a large impact on Patient engagement and disposition [19]. Further studies to assess phy- sician and patient satisfaction as well as medical legal risk could further show the impact of shared decision-making in this patient population.

Limitations

The study was designed around a practice change in a community setting implementing a shared decision-making tool to integrate an ADP and the HEART score. As such, the study has several limitations worth noting. First, using controls from the pre-implementation period

is susceptible to confounding and bias from changing time trend in dis- position outcomes, concurrent interventions in the hospital and differ- ences in patient characteristics. However, it is unethical to have concurrent controls who would be deprived of the opportunity to re- ceive the intervention when the HEART score has proven efficacy. Fur- thermore, when judging from plot of time trend in Fig. 4, the slope of change appeared greater in the post-implementation than in the pre- implementation period indicating some effect attributable to the inter- vention. Finally, although there were some measured disparities of pa- tient characteristics between the two time periods, they did not seem to have a major impact on the results, because the magnitude of associ- ations did not appreciably change after the statistical adjustment. How- ever, we could not exclude unmeasured confounders.

Second, inherent differences and incompatibility between ICD-9 codes in 2014 and ICD-10 codes in 2016 may cause some incomparabil- ity between the pre- and post-patient cohorts. Coding inaccuracy in ICD-9 and ICD-10 billing diagnoses could also result in some misclassi- fication of outcomes. However, we have tried to enhance the specificity of the outcome by supplementing ICD9/10 information with a troponin lab test signifying prime consideration of acute coronary syndrome.

Table 4

Outcomes for patients who were discharged directly from the ED.

Outcome	Pre (%, n)	Post (%, n)	Unadjusted p	Adjusted OR (95% CI)a	p
All patients	3401	3109
All-cause ED bounce back	14.4% (489)	16.5% (514)	0.02	1.14 (0.99, 1.31)	0.07
Cardiac adverse events NSTEMI or STEMI	0.03% (1)	0.13% (4)	0.20	NA	NA
PTCI/PCI	0% (0)	0.06% (2)	0.23	NA	NA
Cardiac arrest	0.03% (1)	0% (0)	0.99	NA	NA
All-cause death	0.76% (26)	0.19% (6)	0.001	0.35 (0.14, 0.88)	0.03
Cardiac events + death	0.79% (27)	0.32% (10)	0.01	0.56 (0.26, 1.18)	0.12
Age b 45 yrs	974	1210
All-cause ED bounce back	12.4% (121)	14.2% (172)	0.22	1.18 (0.90, 1.53)	0.23
Cardiac adverse events NSTEMI or STEMI	0	0	NA	NA	NA
PTCI/PCI	0	0	NA	NA	NA
Cardiac arrest	0	0	NA	NA	NA
All-cause death	0.1% (1)	0.08% (1)	0.99	NA	NA
Cardiac events + death	0.1% (1)	0.08% (1)	0.99	NA	NA
45 yrs b age b 65 yrs	1290	1270
All-cause ED bounce back	14.7% (189)	16.4% (208)	0.23	1.04 (0.83, 1.30)	0.72
Cardiac adverse events NSTEMI or STEMI	0.08% (1)	0.24% (3)	0.37	NA	NA
PTCI/PCI	0% (0)	0.16% (2)	0.25	NA	NA
Cardiac arrest	0	0	NA	NA	NA
All-cause death	0.70% (9)	0.24% (3)	0.09	0.31 (0.07, 1.35)	0.12
Cardiac events + death	0.78% (10)	0.47% (6)	0.33	0.66 (0.22, 1.95)	0.45
Age N 65 yrs	1137	629
All-cause ED bounce back	15.7% (179)	21.3% (134)	0.003	1.24 (0.96, 1.61)	0.10
Cardiac adverse events NSTEMI or STEMI	0	0.16% (1)	0.36	NA	NA
PTCI/PCI	0	0	NA	NA	NA
Cardiac arrest	0.09% (1)	0	0.99	NA	NA
All-cause death	1.41% (16)	0.32% (2)	0.03	0.25 (0.06, 1.12)	0.07
Cardiac events + death	1.41% (16)	0.48% (3)	0.07	0.38 (0.11, 1.33)	0.13

NA indicates that statistics were not performed due to none or too few events.

^a Multivariate logistic regression were adjusted for the same variables as in the Table 2.

Fig. 5. Grouped bar chart comparing patient disposition between different HEART Score risk groups.

Third, the follow-up for ED bounce back and MACE may be incom- plete if the patient was lost to follow-up. This is not an uncommon lim- itation for research that is based on hospital data. However, since most of the covariates in the logistic regression model were presumed corre- lates of propensity for loss to follow-up, we have implicitly controlled for this loss to follow-up bias to some extent. Furthermore, our results are consistent with prior RCT studies that did not find an increase in ED bounce back or MACE rates with more robust follow-up [7].

Lastly, we were unable to ascertain the compliance of the protocol. Though we believe clinicians were using the HEART score as part of their chest pain evaluation, there was a subset of total patients who had the HEART score documented in the chart. In particular, there were a low number of high-risk patients documented. Many of these high-risk patients may have had a positive troponin test or appeared to be clearly suffering from acute coronary syndrome so the HEART score may not have been useful in the Disposition decision. Previous studies found a N20% non-adherence rate so we made the assumption that there would be some non-compliance [26,27]. In this study, we ex- amined the global impact of this protocol among all chest pain patients presenting to the ED, not just those with documented HEART scores. This impact has naturally incorporated the effect of non-compliance to reflect a more realistic setting where non-compliance is unpreventable.

Conclusions

The HAS-Choice pathway suggests that implementation of the HEART score, an ADP, and shared decision-making using a visual aid, is feasible to achieve and can safely decrease hospital admissions. This study demonstrates generalizability to a larger diverse population of pa- tients in a community emergency department setting.

Funding source

St. Joseph Mercy Ann Arbor Research Committee internal funding.

Presentations

Gafni-Pappas G, DeMeester S, Boyd M, Ganti A. “shared decision making Employing HEART Score and a Visual Aid in Patients Presenting with Chest Pain to a Community Emergency Department.” Spotlight Oral Presentation. Society for Academic Emergency Medicine Annual Meeting. New Orleans 2016.

Acknowledgement

We would like to thank the staff in the St. Joseph Mercy Hospital Ac- ademic Research Department for their regulatory and administrative

support of this study. We are also indebted to the staff at the Quality In- stitute for their data extraction.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2018.02.005.

References

1. Amsterdam EA, Kirk JD, Bluemke DA, et al. Testing of low-risk patients presenting to the emergency department with chest pain: a scientific statement from the American Heart Association. Circulation 2010;122(17):1756-76.
2. Penumetsa SC, Mallidi J, Friderici JL, Hiser W, Rothberg MB. Outcomes of patients ad- mitted for observation of chest pain. Arch Intern Med 2012;172(11).
3. Pope JH, Aufderheide TP, Ruthazer R, et al. missed diagnoses of acute cardiac ische- mia in the emergency department. N Engl J Med 2000;342(16):1163-70.
4. Pines JM, Isserman JA, Szyld D, Dean AJ, McCusker CM, Hollander JE. The effect of physician risk tolerance and the presence of an observation unit on decision making for ED patients with chest pain. Am J Emerg Med 2010;28(7):771-9.
5. Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and stroke statistics-2012 up- date: a report from the American heart Association. Circulation 2012;125(1).
6. Field JM, Hazinski MF, Sayre MR, et al. Part 1: executive summary: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency car- diovascular care. Circulation 2010;122(Suppl. 3).
7. Mahler SA, Riley RF, Hiestand BC, et al. The HEART pathway randomized trial: iden- tifying emergency department patients with acute chest pain for early discharge. Circ Cardiovasc Qual Outcomes 2015;8(2):195-203.
8. Long B, Oliver J, Streitz M, Koyfman A. An end-user’s guide to the HEART score and pathway. Am J Emerg Med 2017;35(9):1350-5.
9. Leite L, Baptista R, Leitao J, et al. Chest pain in the emergency department: risk strat- ification with Manchester triage system and HEART score. BMC Cardiovasc Disord 2015;15(1):48.
10. Cullen LA, Mills NL, Mahler S, Body R. Early rule-out and rule-in strategies for myo- cardial infarction. Clin Chem 2017;63(1):129-39.
11. Callaway CW, Donnino MW, Fink EL, et al. Part 8: post-cardiac arrest care: 2015 American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2015;132(18):S465-82.
12. Six a J, Backus BE, Kelder JC. Chest pain in the emergency room: value of the HEART score. Neth Hear J 2008;16(6):191-6.
13. Backus BE, Six AJ, Kelder JC, et al. A prospective validation of the HEART score for chest pain patients at the emergency department. Int J Cardiol 2013;168(3):2153-8.
14. Six AJ, Cullen L, Backus BE, et al. The HEART score for the assessment of patients with chest pain in the emergency department: a multinational validation study. Crit Pathw Cardiol 2013;12(3):121-6.
15. Fanaroff AC, Rymer JA, Goldstein SA, Simel DL, Newby LK. Does this patient with chest pain have acute coronary syndrome?: the rational clinical examination sys- tematic review. JAMA 2015;314(18):1955-65.
16. Stopyra JP, Miller CD, Hiestand BC, et al. Chest pain risk stratification: a comparison of the 2-hour accelerated diagnostic protocol (ADAPT) and the HEART pathway. Crit Pathw Cardiol 2016;15(2):46-9.
17. Backus BE, Six AJ, Kelder JH, Gibler WB, Moll FL, Doevendans PA. Risk scores for pa- tients with chest pain: evaluation in the emergency department. Curr Cardiol Rev 2011;7(1):2-8.
18. Mahler SA, Miller CD, Hollander JE, et al. Identifying patients for early discharge: per- formance of decision rules among patients with acute chest pain. Int J Cardiol 2013; 168(2):795-802.
19. Hess EP, Hollander JE, Schaffer JT, et al. Shared decision making in patients with low risk chest pain: prospective randomized pragmatic trial. BMJ 2016:i6165.
20. Hess EP, Knoedler MA, Shah ND, et al. The chest pain choice decision aid: a random- ized trial. Circ Cardiovasc Qual Outcomes 2012;5(3):251-9.
21. Cuckler GA, Sisko AM, Keehan SP, et al. National health expenditure projections, 2012-22: slow growth until coverage expands and economy improves. Health Aff 2013;32(10):1820-31.
22. Martin AB, Hartman M, Washington B, Catlin A. National health spending: faster growth in 2015 as coverage expands and utilization increases. Health Aff 2017;36 (1):166-76.
23. Flynn D, Knoedler MA, Hess EP, et al. Engaging patients in health Care decisions in the emergency department through shared decision-making: a systematic review. Acad Emerg Med 2012;19(8):959-67.
24. Melki D, Jernberg T. HEART Score. Crit Pathw Cardiol 2013;12(3):127-31.
25. Riley RF, Miller CD, Russell GB, et al. cost analysis of the history, ECG, age, risk fac- tors, and initial troponin (HEART) pathway randomized control trial. Am J Emerg Med 2017;35(1):77-81.
26. Mahler SA, Riley RF, Russell GB, et al. Adherence to an accelerated diagnostic proto- col for chest pain: Secondary analysis of the HEART pathway randomized trial. Aca- demic emergency medicine; 2016. p. 70-7.
27. Than M, Aldous S, Lord SJ, et al. A 2-hour diagnostic protocol for possible cardiac chest pain in the emergency department. JAMA Intern Med 2014;174(1):51.