CARDIOLOGY/ORIGINAL RESEARCH

An Analysis of the Association of Society of Chest Pain CentersAccreditation to American College of Cardiology/American Heart

Association Non–ST-Segment Elevation Myocardial InfarctionGuideline Adherence

Abhinav Chandra, MDSeth W. Glickman, MD, MBAFang-Shu Ou, MSW. Frank Peacock, MDJames K. McCord, MDCharles B. Cairns, MDEric D. Peterson, MD, MPHE. Magnus Ohman, MDW. Brian Gibler, MDMatthew T. Roe, MD, MHS

From the Duke University Medical Center, Durham, NC (Chandra, Glickman); the Duke ClinicalResearch Institute, Durham, NC (Ou, Peterson, Ohman, Roe); The Cleveland Clinic, Cleveland, OH(Peacock); the Henry Ford Heart and Vascular Institute, Detroit, MI (McCord); University of NorthCarolina, Chapel Hill, NC (Cairns); and the University of Cincinnati, College of Medicine, Cincinnati,OH (Gibler).

Study objective: Since 2003, the Society of Chest Pain Centers (SCPC) has provided hospital accreditation foracute coronary syndrome care processes. Our objective is to evaluate the association between SCPCaccreditation and adherence to the American College of Cardiology/American Heart Association (ACC/AHA)evidence-based guidelines for non–ST-segment elevation myocardial infarction (NSTEMI). The secondaryobjective is to describe the clinical outcomes and the association with accreditation.

Methods: We conducted a secondary analysis of data from patients with NSTEMI enrolled in the Can Rapid RiskStratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines (CRUSADE) quality improvement initiative in 2005. The analysis explored differences betweenSCPC-accredited and nonaccredited hospitals in evidence-based therapy given within the first 24 hours (includingaspirin, �-blocker, glycoprotein IIb/IIIa inhibitors, heparin, and ECG within 10 minutes).

Results: Of 33,238 patients treated at 21 accredited hospitals and 323 nonaccredited hospitals, those atSCPC-accredited centers (n�3,059) were more likely to receive aspirin (98.1% versus 95.8%; odds ratio [OR]1.73; 95% confidence interval [CI] 1.06 to 2.83) and �-blockers (93.4% versus 90.6%; OR 1.68; 95% CI 1.04 to2.70) within 24 hours than patients at non-SCPC-accredited centers (n�30,179). No difference was observed inobtaining a timely ECG (40.4% versus 35.2%; OR 1.28; 95% CI 0.98 to 1.67), administering a glycoprotein IIb/IIIa inhibitor (OR 1.30; 95% CI 0.93 to 1.80), or administering heparin (OR 1.12; 95% CI 0.74 to 1.70). Also,there was no significant difference in risk-adjusted mortality for patients treated at SCPC hospitals versusnonaccredited hospitals (3.4% versus 3.5%; adjusted OR 1.17; 95% CI 0.88 to 1.55).

Conclusion: SCPC-accredited hospitals had higher NSTEMI ACC/AHA evidence-based guideline adherence in thefirst 24 hours of care on 2 of the 5 measures. No difference in outcomes was observed. Further studies areneeded to better understand the association between SCPC accreditation and improved care for patients withacute coronary syndrome. [Ann Emerg Med. 2009;54:17-25.]

Provide feedback on this article at the journal’s Web site, www.annemergmed.com.

0196-0644/$-see front matterCopyright © 2008 by the American College of Emergency Physicians.doi:10.1016/j.annemergmed.2009.01.025

INTRODUCTIONBackground

Emergency departments (EDs) in the United States treated

more than 115 million patients in 2005, including more than 5

Volume , . : July

million with undifferentiated acute chest pain.1 Emergencyphysicians are responsible for rapidly identifying and initiatingevidence-based treatment in patients with acute coronary

syndromes. The American College of Cardiology/American

Annals of Emergency Medicine 17

SCPC Accreditation and NSTEMI Guidelines Adherence Chandra et al

Heart Association (ACC/AHA) publishes evidence-basedguidelines for the management of non–ST-segment elevationmyocardial infarction (NSTEMI).2 The ACC/AHA guidelinesrecommend that patient risk stratification be performed byconsideration of history, physical examination, and an ECG. Ifearly diagnostics are inconclusive, serial ECGs and serial cardiacmarker tests may be indicated. According to this information,patients are stratified into ST-segment elevation myocardialinfarction (STEMI), NSTEMI, unstable angina, or noncardiacchest pain cohorts, and appropriate therapy is initiated.

ImportanceDespite widespread dissemination of the ACC/AHA

guidelines, adherence has been inconsistent.3,4 The Can RapidRisk Stratification of Unstable Angina Patients SuppressAdverse Outcomes With Early Implementation of the AmericanCollege of Cardiology/American Heart Association Guidelines(CRUSADE) initiative and the Society of Chest Pain Centers(SCPC) were formed to help in the care of this population andimprove adherence. The CRUSADE initiative was establishedto improve adherence with ACC/AHA recommendations byproviding feedback to each member site on adherence rates andcomparison rates. The SCPC was formed in 1998 to educatephysicians and assist hospitals in the development ofprotocolized, evidence-based risk stratification and treatmentpathways and protocols for patients with suspected acute

Editor’s Capsule Summary

What is already known on this topicThe association of chest pain center accreditationwith core measures of care for acute myocardialinfarction remains uncertain.

What question this study addressedThis large retrospective study questioned whetherSociety of Chest Pain Center accreditation wasassociated with better adherence to guidelines foremergency myocardial infarction care and betterclinical outcomes.

What this study adds to our knowledgeChest pain center accreditation was associated witha small improvement in adherence to clinicalguidelines but no improvement in major clinicaloutcomes.

How this might change clinical practiceThis work clarifies the potential benefit of the chestpain center accreditation process and might helpguide whether this would be a worthwhileundertaking.

coronary syndrome. These pathways incorporate many of the

18 Annals of Emergency Medicine

ACC/AHA recommendations, thus helping with evidence-basedcare. In 2003, the SCPC initiated an accreditation process basedon 8 key principles to evaluate the triage and diagnosticprocesses. To date, approximately 440 centers have beenaccredited or are in the process of being accredited by theSCPC.5 The association of accreditation with adherence to theACC/AHA recommendations for patients with NSTEMI hasnot been evaluated.

Goals of This InvestigationOur primary objective was to evaluate the association

between SCPC accreditation and adherence to the ACC/AHAevidence-based guidelines for NSTEMI. A secondary analysiswas performed to describe the association between accreditationand patient outcomes.

MATERIALS AND METHODSStudy Design

A secondary analysis of the CRUSADE initiative studydesign was performed to assess the association between SCPCaccreditation and adherence to treatment guidelines and patientoutcomes.

The CRUSADE initiative is a voluntary, observational,quality improvement initiative begun on January 1, 2001, thatwas designed to improve the quality of evidence-based care forpatients with NSTEMI and acute coronary syndrome.6,7 Inconfirmed NSTEMI patients, participating hospitals were askedto collect and submit clinical information about inhospital careand outcomes on all consecutive patients with a confirmeddiagnosis of NSTEMI. To be eligible, patients were required topresent at a CRUSADE hospital within 24 hours of ischemicsymptoms lasting at least 10 minutes and in combination witheither positive cardiac biomarker results (troponin or creatinekinase-MB) or ischemic ST-segment electrocardiographicchanges (ST-segment depression or transient ST-segmentelevation).

Trained data collectors at each hospital used standardizeddefinitions and forms for retrospective data abstraction.Variables included demographic characteristics, clinicalpresentation, medical history, treatments administered,associated major contraindications to evidence-based therapies,inhospital outcomes, discharge recommendations, andinterventions. Multiple procedures were used to monitor andensure the quality of the CRUSADE database.6,7 Sites wereencouraged to submit data on all unstable angina/NSTEMIpatients. Data collection was retrospective, primarily identifiedthrough a screening of relevant International Classification ofDiseases, Ninth Revision codes. Sites were sometimes limitedfrom entering all patients for 2 main reasons: (1) resources werelimited to pay hospital personnel to abstract data; and (2)medical records could not be obtained in time to meet thequarterly data submission deadlines. To ensure that sites werenot censoring certain types of patients (ie, those with pooreroutcomes), coordinating center personnel reviewed site

mortality rates against what would be expected from the case

Volume , . : July

Chandra et al SCPC Accreditation and NSTEMI Guidelines Adherence

mix of patients (on a quarterly basis). Where these ratesdeviated, site personnel were called to help provide guidance onlocal data submission processes.

A variety of performance feedback and quality improvementinterventions was provided to sites, including reports thatbenchmarked against national standards and other CRUSADEhospitals, as well as educational interventions, such as treatmentalgorithms and risk-stratification tools. Participating institutionswere required to comply with their local regulatory and privacyguidelines and to submit the CRUSADE protocol for reviewand approval by their institutional review boards. Because datawere used primarily at the local sites for quality improvement,all sites were granted a waiver of informed consent under theCommon Rule.

Selection of ParticipantsWe selected 2005 for our analysis because this was the first

year with a sufficient number of SCPC-accredited hospitals andpatients to perform a meaningful analysis. SCPC accreditationlasts for 3 years; thus, all centers accredited in 2003 and 2004were still accredited in 2005 and eligible for inclusion in thisanalysis. Centers from 2006 were not entered because some mayhave been in the act of recertifying, and the data in 2006 maynot have represented their routine care. The time required torecertify ranges from weeks to months. Thus, any data includedfrom centers being recertified may have been biased and nottruly representative of standard care provided at that site.Additionally, in 2006, some of the hospitals’ accreditationexpired and their status changed to nonaccredited. If 2006 datawere added to this analysis, the handling of hospitals withexpired accreditation would become an issue. We do not believethat a hospital with an expired accreditation means the hospitalwill experience a decrease in quality of care; however, it is alsonot reasonable to keep it in the accredited group when we try toexplore the association between accreditation status and qualityof care. Thus, to preserve the quality of the data andconclusions, we limited our analysis to 2005.

We identified 364 total hospitals in CRUSADE, of which 22were accredited by the SCPC and 342 were nonaccredited.Twenty hospitals enrolling fewer than 10 patients in 2005 wereexcluded from the analysis, 1 from the SCPC-accredited groupand 19 from the nonaccredited group. We used a generalizedestimating equation to control for within-hospital clustering.When there are only a few observations within each hospital, itwill affect mean estimation and power of the testing. Removinghospitals with fewer than 10 patients is expected to improveefficiency of the testing and the estimation. After exclusion oflow-enrolling sites, 7 SCPC-accredited centers from 2003, 14SCPC-accredited hospitals from 2004, and 323 nonaccreditedsites were incorporated into the final analysis (Figure).

After hospitals were narrowed for analysis, patient inclusionwas also evaluated. First, patients with negative cardiacbiomarker results were excluded, although labeled as havingunstable angina. The label of unstable angina was placed

retrospectively, and thus the initial physician may not have

Volume , . : July

recognized it and may not have initiated guideline care.Therefore, to make the data more interpretable and meaningful,we excluded this group. Additionally, patients were excludedwho died within 24 hours because the hospital did not have afull 24 hours to comply with the ACC/AHA guidelines. Last,patients transferred into CRUSADE hospitals were excludedfrom the analyses of ECG within 10 minutes. Patientstransferred out of CRUSADE hospitals were excluded from theanalyses of inhospital clinical events.

Five acute-care process measures designated as class I (usefuland effective) in the ACC/AHA NSTEMI guidelines wereevaluated. These included performing an ECG within 10minutes of ED presentation, and administration of aspirin, �-blocker, unfractionated or low-molecular-weight heparin, andglycoprotein IIb/IIIa inhibitors within 24 hours ofhospitalization. For each specific acute therapy, patients wereexcluded from the final analyses if they had documentedcontraindications to that intervention (eg, aspirin allergy).6,7

Inhospital mortality was defined as death from any cause duringa patient’s hospitalization.

We also report postadmission infarction. In patients withNSTEMI before revascularization, this was defined as a new qwave in at least 2 contiguous leads or an increase in biomarkersby at least 50% above the most recent values. In patients whohad revascularization within 24 hours, we definedpostadmission infarction as an increase in creatine kinase-MBby at least 50% over the most recent value, an increase increatine kinase-MB to a value at least 3 times the upper limit ofnormal, or a new q wave in 2 contiguous leads. Last, we definedpostadmission infarction within 24 hours of coronary arterybypass grafting as an increase in creatine kinase-MB 5 timesupper limit of normal or a new q wave in at least 2 contiguousleads.

Sensitivity AnalysesBaseline characteristics, treatment profiles, procedure use,

and clinical outcomes were summarized by hospital accreditedstatus. Continuous variables are presented as medians withinterquartile percentiles, and categorical variables are expressedas percentages. Overall missing data percentage is less then 2%,except body mass index (BMI) which has 5% missing values.Missing values of age, pulse rate, and systolic blood pressurewere imputed to median. Missing values of BMI were imputedto sex-specific median. Missing values of the remaining riskfactors were defaulted to their most common value. There is nomissing value for hospital-level variables, number of hospitalbeds, region, teaching status, and facility type. Missing values ofclinical events were excluded when the clinical event was used asanalysis outcome. Because the relationship between thecontinuous variables and the outcome is probably nonlinear,continuous variables were fitted by using linear splines inmultivariable logistic models.

Multivariate logistic regression analyses using generalizedestimating equations methods with exchangeable working

correlation structure determined whether SCPC accreditation

Annals of Emergency Medicine 19

from

SCPC Accreditation and NSTEMI Guidelines Adherence Chandra et al

was associated with the use of ACC/AHA class I therapies ordeath.8 The generalized estimating equation method was usedto account for within-hospital clustering because patients at thesame hospital are more likely to have similar responses relativeto patients in other hospitals (ie, within-center correlation forresponse). The method produces estimates similar to those fromordinary logistic regression, but the estimated variances of theestimates are adjusted for the correlation of outcomes withineach hospital.8 Models were adjusted for age, sex, race, BMI,family history of coronary artery disease, hypertension, diabetesmellitus, current/recent smoker (within 6 weeks ofhospitalization), hypercholesterolemia, previous myocardialinfarction, previous percutaneous coronary intervention,previous coronary artery bypass grafting, previous congestiveheart failure, previous stroke, renal insufficiency (serumcreatinine level �2.0 mg/dL, calculated creatinine clearance�30 mL/min, or need for renal dialysis), electrocardiographic

Figure. Patient selection

findings at admission, signs of congestive heart failure, pulse

20 Annals of Emergency Medicine

rate, and systolic blood pressure at admission. Models were alsoadjusted for nonclinical variables. These potentiallyconfounding variables are primary attending physician,insurance status, number of hospital beds, region, teachingstatus, and facility type. Odds ratios (ORs) and 95% confidenceintervals (CIs) were calculated to determine the associationbetween SCPC accreditation status and quality of care. � Wasdefined as a value less than 0.05. Analyses were performed usingSAS software (version 9.0, SAS Institute, Inc., Cary, NC).

RESULTSThe registry contained a total of 189,065 entries at analysis.

The number was narrowed to 35,141 patients after limiting ofthe analysis to 2005. Patients were excluded from this analysisfor several reasons. First, 1,646 were excluded because ofnegative cardiac biomarker results. An additional 174 wereexcluded because of death within 24 hours because the hospital

the CRUSADE Initiative.

did not have a full 24 hours to comply with the ACC/AHA

Volume , . : July

Chandra et al SCPC Accreditation and NSTEMI Guidelines Adherence

guidelines, leaving 33,321 patients. Finally, 83 patients fromhospitals treating fewer than 10 patients in 2005 were excludedto provide more stable estimates of performance measures.Specifically, 3 patients were excluded from 1 SCPC-accreditedhospital and 80 were excluded from 19 nonaccredited hospitals.Thus, after exclusion of a total of 1,903 patients, 33,238patients from 344 CRUSADE hospitals formed the analysiscohort. Of these, 3,059 (9.2%) patients were treated at one of21 SCPC-accredited chest pain hospitals, and the remaining30,179 (90.8%) were treated at one of 323 non–SCPC-accredited hospitals (Figure 1).

A total of 344 hospitals were included in this analysis. Ofthese, 21 were SCPC accredited and 323 were not. Six (28.6%)SCPC-accredited sites were academic compared with 81(25.1%) of non–SCPC-accredited sites (Table 1). Of theSCPC-accredited hospitals, 18 (85.7%) provided diagnosticcardiac catheterization, therapeutic cardiac catheterization, andbypass grafting surgery compared with 221 (68.4%) of thenonaccredited sites. More than 95% of the SCPC-accreditedhospitals were located in the Midwest and the South comparedwith only 60% of the non-SCPC centers. The SCPC-accreditedcenters tended to have more hospital beds (median 450 [rangeof 339-621] versus a median of 309 [range of 204-453] beds).

Characteristics of Study SubjectsThere were small differences in patient characteristics

between the 2 hospital groups. Patients at SCPC-accreditedhospitals were younger and more likely to be of a minority race.Additionally, patients at SCPC-accredited sites had a higherBMI and were more likely to have a family history of coronaryartery disease and some cardiovascular risk factors (Table 2).Some similarities were observed in the clinical presentation ofthe patients in the 2 groups, including similar presenting pulserates and systolic blood pressure. Differences were observed at

Table 1. Hospital characteristics.

CharacteristicsNonaccredited

CentersAccredited

Centers

No. 323 21Academic center, No. (%)* 81 (25) 6 (29)Cardiac facilities, No. (%)No catheterization 34 (11) 0Catheterization only 39 (12) 2 (10)PCI only, no surgery 29 (9) 1 (5)PCI and surgery 221 (68) 18 (86)Region, No. (%)West 42 (13) 0Northeast 87 (27) 1 (5)Midwest 79 (24) 9 (43)South 115 (36) 11 (52)Number of beds

†309 (204, 453) 450 (339, 621)

PCI, Percutaneous coronary intervention.*A member of the Council of Teaching Hospitals.†Median values with 25th, 75th percentiles displayed for continuous variables.

presentation in the presence of signs of congestive heart failure

Volume , . : July

(24.2% at non–SCPC-accredited sites versus 21.0% at SCPC-accredited sites).

SCPC-accredited hospitals were more likely to administerACC/AHA guideline-recommended therapies for NSTEMI,including aspirin (98.1% versus 95.8%; adjusted OR 1.73; 95%CI 1.06 to 2.83) and �-blockers (93.4% versus 90.6%; adjustedOR 1.68; 95% CI 1.04 to 2.70) (Tables 3 and 4). No differencewas observed for patients receiving an ECG within 10 minutesof presentation, glycoprotein IIb/IIIa inhibitors, or heparinamong patients from SCPC-accredited hospitals versus patientsfrom nonaccredited hospitals.

Accredited hospitals performed more diagnosticcatheterizations (Table 5) and more procedures within the first2 days than non–SCPC-accredited sites. In addition todiagnostic procedures, a greater number of patients receivedpercutaneous coronary intervention, more of which wereperformed within 48 hours at accredited centers compared withthe nonaccredited hospitals. No difference was observed in therates of coronary artery bypass grafting between the 2 groups.

Unadjusted analysis of clinical events during thehospitalization was also compared. No differences were observedin mortality, reinfarction, cardiogenic shock, ischemic stroke,hemorrhagic stroke, congestive heart failure, or overall majorbleeding episodes (Tables 3 and 4).

LIMITATIONSThere are several issues that should be considered in the

interpretation of the results of this study. First, this study is aretrospective observational one and does not establish a causalrelationship between SCPC accreditation and guidelineadherence. Further studies are required, including prospectivelongitudinal assessments of care at these hospitals. Second,although the CRUSADE registry is among the largest acutecoronary syndrome registries in the world, it enrolled only asubset of the total non-ST segment elevation acute coronarysyndrome population admitted to US hospitals each year, so aselection bias related to the unequal geographic distribution ofparticipating hospitals may have influenced the results. Selectionbias could also occur if hospitals did not report patientinformation on selected cases (ie, patients with poor outcomes),although CRUSADE personnel closely monitored and routinelyprovided guidance to hospitals on data submission processes.Third, our evaluation did not analyze low-risk patients,including those without positive cardiac biomarker results.Fourth, we are unable to evaluate clinical outcomes, proceduralutilization, and discharge treatment patterns in patientstransferred to other institutions because of current privacyregulations that restrict anonymous data collection after hospitaltransfer. Fourth, data cannot be adjusted for undocumentedcontraindications to guideline recommendations that are notdetailed in medical records but that may lead to appropriatewithholding of certain medications. Fifth, discrepancies inresources available to facilitate documentation may differbetween the 2 groups; however, participation in the

accreditation process did not depend on resources to document

Annals of Emergency Medicine 21

SCPC Accreditation and NSTEMI Guidelines Adherence Chandra et al

Table 2. Patient characteristics.

CharacteristicsOverall Results(N�33,238)*

Nonaccredited Centers(N�30,179)*

Accredited Centers(N�3,059)*

DemographicsAge, y

†68 (56, 79) 68 (56, 79) 65 (55, 78)

BMI†

28 (25.0, 32.2) 27.9 (24.5, 32.2) 28.3 (25.0, 32.6)Sex, male 60.3 60.2 61.3Race

White 80.7 80.8 80.0Black 10.1 9.9 11.5Asian 1.1 1.1 0.6Hispanic 4.0 4.0 3.6Other 2.0 2.0 2.0

Medical historyFamily Hx of CAD 32.8 32.6 34.8Hx hypertension 71.1 71.1 71.1Diabetes mellitus 33.1 33.3 31.2Current/recent smoker 27.1 26.7 31.2Dyslipidemia 52.3 52.1 54.1Previous MI 27.8 27.7 29.3Previous CHF 16.8 17.0 14.5Previous stroke 10.0 10.0 9.6Previous CABG 18.3 18.3 18.5Previous PCI 20.8 20.3 25.0Renal insufficiency 13.6 10.8Signs/symptoms on presentationSigns of CHF 23.9 24.2 21.0Systolic BP, mm Hg

†144 (123, 164) 143 (123, 164) 145 (125, 165)

HR, beats per minute†

84 (71, 100) 84 (71, 100) 83 (70, 98)

Hx, History; CAD, coronary artery disease; MI, myocardial infarction; CHF, congestive heart failure; CABG, coronary artery bypass grafting; BP, blood pressure; HR,pulse rate.*For categorical variables, percentages are reported.†

Median values with 25th, 75th percentiles are displayed for continuous variables.

Table 3. Unadjusted analysis of process of care measures.

OutcomeOverall Results

(N�33,238)Nonaccredited Centers

(N�30,179)Accredited Centers

(N�3,059)

Acute care measuresArrival to first ECG, min* 14 (7, 32) 15 (7, 33) 12 (7, 24)ECG within 10 min of arrival, % 35.6 35.2 40.4Aspirin, % 96.0 95.8 98.1Any heparin, % 88.2 87.9 90.6Heparin: IV unfractionated, % 51.9 51.2 58.5Heparin: low molecular weight, % 43.2 43.5 40.7�-Blocker, % 90.9 90.6 93.4GP IIb/IIIa inhibitor, % 48.0 46.6 60.3ACE inhibitor, % 48.8 48.2 54.1Clopidogrel, % 56.2 55.3 65.1Inhospital clinical eventsPostadmission infarction, % 1.9 2.0 1.7Cardiogenic shock, % 2.2 2.2 1.8CHF, % 7.4 7.6 6.2Death, % 3.5 3.5 3.4Death or MI, % 5.0 5.0 4.9Major bleeding (non-CABG patients), % 10.8 11.0 9.5

IV, Intravenous; GP, glycoprotein; ACE, angiotensin-converting enzyme.

*Median values with 25th, 75th percentiles displayed for continuous variables.

22 Annals of Emergency Medicine Volume , . : July

Chandra et al SCPC Accreditation and NSTEMI Guidelines Adherence

core measures. Last, it is not possible to adjust for or identifyintentional omissions of patients whose care did not meetquality standards. However, these processes are not affected bySCPC accreditation; thus, the discrepancies, if present, areexpected to be uniform across all sites.

DISCUSSIONAccreditation by the SCPC is associated with improved

ACC/AHA guideline adherence to aspirin and �-blockeradministration and no difference in door-to-ECG times orheparin and glycoprotein IIb/IIIa administration, as evaluatedby the CRUSADE registry of NSTEMI patients. Although ouranalysis did not demonstrate a difference in death, myocardialinfarction, cardiogenic shock, postadmission infarction,congestive heart failure, or major bleeding, larger studies havedemonstrated that better guideline adherence is associated withlower mortality rates.3

Ischemic heart disease remains the leading cause of deathworldwide.9 The burden of treating the approximately 6million patients seeking evaluation for acute chest painannually in the United States is large, resulting in intensiveresource utilization and ED crowding.10-14 The SCPC wasone of the first organizations attempting to standardize careand evidence-based treatment for all ED patients with acute,undifferentiated chest pain (including those at low risk foracute coronary syndrome to STEMI). The SCPC focusesprimarily on processes of care in the ED. Additionally, theorganization helps to structure the transfer of care from theED to the inpatient setting; however, its accreditationprocess does not focus on inpatient care. The present study isthe first attempt to evaluate the effectiveness of SCPC effortsin the care of the ED patients with NSTEMI.

We examined whether hospitals with SCPC accreditation

Table 4. Outcome-adjusted ORs associated with chest paincenter accreditation.

Outcome Adjusted OR (95% CI)

Acute aspirin*†

1.73 (1.06-2.83)Acute �-blocker*

†1.68 (1.04-2.70)

Acute GP IIb/IIIa inhibitor*†

1.30 (0.93-1.80)Acute heparin*

†1.12 (0.74-1.70)

ECG in 10 min*†

1.28 (0.98-1.67)Cath within 24 h* 1.24 (0.93-1.66)Postadmission infarction* 0.70 (0.36-1.38)Cardiogenic shock* 0.84 (0.47-1.50)CHF* 0.96 (0.41-2.28)Death* 1.17 (0.88-1.55)Death/MI* 1.07 (0.80-1.42)Major bleeding (non-CABG patients)* 0.88 (0.67-1.14)

Cath, Catheterization.*Variables in the model: age, male sex, white race, BMI, pulse rate, systolicblood pressure, previous percutaneous coronary intervention, history of renalinsufficiency, previous CHF, diabetes mellitus, CHF at presentation, ST depres-sion.†Additional variables in the model: primary care by cardiologist, insurance, num-ber of hospital beds, region, hospital facility, teaching hospital.

were more likely to administer ACC/AHA guidelines-

Volume , . : July

recommended therapy for patients with NSTEMI. Ourfindings were statistically significant; however, the differenceswere clinically small. For 2 of the 5 process measures studied(acute aspirin and �-blocker therapy), SCPC-accreditedhospitals demonstrated higher levels of adherence to thesetherapies. No benefit was observed in obtaining an ECGwithin 10 minutes or administering heparin or glycoproteinIIb/IIIa inhibitors between the 2 groups. The reason forthese differences is unclear. The first possible explanation isthat the accreditation process actually helps hospitalsimprove patient care. We questioned the difference noted for�-blocker administration because more patients presentedwith symptoms of congestive heart failure to non–SCPC-accredited hospitals (Table 2). After adjusting for thepresence of congestive heart failure, �-blocker administrationrates were still statistically significant in the accredited group;thus, the rates of congestive heart failure at presentation werenot the reason for the difference.

On the other hand, another explanation may be that noimprovement in care occurred and the trends observed may be aresult of better documentation of administration, or they maybe due to better documentation of contraindications toadministration. Last, this finding may be a statistical anomalycaused by a large sample size with little clinical importance.

In addition to understanding the potential reasons for thedifference, we need to place the findings in the context of thecurrent evidence-based medicine. Specifically, theClopidrogel and Metoprolol in Myocardial Infarction(COMMIT) trial15 was conducted to identify superiorityand evaluated the effect of �-blockers in patients withmyocardial infarction and the link to cardiogenic shock.These and other trials have now de-emphasized the role of�-blockers acutely. Thus, the difference observed in rates of�-blocker administration in our analysis was statisticallysignificant but clinically carries less importance in light ofthe COMMIT conclusions.

In regard to inhospital clinical events, no benefit wasobserved in the adjusted rates of congestive heart failure,cardiogenic shock, death, or a combination end point of deathor myocardial infarction among the 2 hospital groups. TheSCPC accreditation process does not directly address inhospitalcare and thus may explain the lack of difference in clinicalevents during the hospitalization. An explanation may be thatthe absolute differences between the accredited andnonaccredited hospitals’ acute care characteristics were clinicallysmall; thus, the subsequent rates of clinical events were alsosmall. Additionally, in this cohort of hospitals, perhaps morethan a few characteristics need to be improved by accreditationbefore a difference in outcomes is realized. Last, because all siteswere part of the CRUSADE initiative to improve care, thisprogram may have resulted in a lack of measurable differences ofinhospital clinical events.

Our findings have been replicated by Ross et al,16 who

reported Centers for Medicare & Medicaid Services core

Annals of Emergency Medicine 23

SCPC Accreditation and NSTEMI Guidelines Adherence Chandra et al

measures in patients with ST-segment elevation and NSTEMIs.Although they demonstrated a higher adherence with acuteaspirin and �-blocker administration as well, they were unableto adjust for variances in hospitals and patients. Furthermore,the cohort across whom they performed their analysis was notuniform in their desire for quality care. These differencesbetween our articles are what make our findings moreimpressive. All the hospitals in the CRUSADE analysisdedicated resources and demonstrated an interest in improvingmanagement of acute coronary syndrome by participating inthis care improvement initiative. Yet, after controlling formultiple variables, the performance rates of 2 of the 5 aspirinand �-blocker administration acute core measures were stilldifferent and unexpected. Hospitals considering accreditation bythe SCPC should review the Ross et al16 results and our resultsand evaluate whether the needs of their institutions parallel thefindings reported by our respective analyses.

We report that accredited members of the SCPC have higheradherence to the ACC/AHA guidelines for administration ofaspirin and �-blockers within 24 hours. To our knowledge, thisis one of the first studies to evaluate the association betweenaccreditation by the SCPC and quality of care for ED patientswith NSTEMI. Future prospective studies are needed tounderstand the effect on care of all patients with chest pain, aswell as the long-term effect of accreditation by the SCPC.

Supervising editor: Keith A. Marill, MD

Author contributions: AC was responsible for study conceptionand design and drafting of the article. F-SO was responsiblefor analysis and interpretation of data. AC, SWG, F-SO, WFP,JKM, CBC, EDP, EMO, WBG, and MTR revised article criticallyfor important intellectual content. AC, SWG, F-SO, WFP, JKM,CBC, EDP, EMO, WBG, and MTR were responsible for finalapproval of the submitted article. AC takes responsibility forthe paper as a whole.

Funding and support: By Annals policy, all authors are requiredto disclose any and all commercial, financial, and otherrelationships in any way related to the subject of this article,that might create any potential conflict of interest. See theManuscript Submission Agreement in this issue for examples

Table 5. Incidence and time to hospital procedures.

Inhospital Procedures, %Overall*

(N�33,238)

Diagnostic catheterization 83.2Catheterization within 24 hours of arrival 47.5Catheterization within 48 hours of arrival 65.8Interventional proceduresPCI 52.3PCI within 24 h of arrival 31.7PCI within 48 h of arrival 42.3CABG 12.4

*For categorical variables, percentages are reported.

of specific conflicts covered by this statement. CRUSADE is

24 Annals of Emergency Medicine

funded by the Schering-Plough Corporation. Bristol-MyersSquibb/Sanofi-Aventis Pharmaceuticals Partnership providesadditional funding support. Millennium Pharmaceuticals, Inc.,also funded this work. WFP and JKM are directly involved withthe Society of Chest Pain Centers. No other conflict of interestexists as related to the topic of this manuscript.

Publication dates: Received for publication August 4, 2008.Revisions received December 19, 2008, and January 12,2009. Accepted for publication January 26, 2009. Availableonline March 12, 2009.

Reprints not available from the authors.

Address for correspondence: Abhinav Chandra, MD, DUMCBox 3096, Durham, NC 27710; 919-416-8201, fax 919-681-8521; E-mail: abhinav.chandra@duke.edu.

REFERENCES1. Nawar EW, Niska RW, Xu J. National Ambulatory Medical Care

Survey: 2005 Emergency Department Summary. Adv Data No.386. Hyattsville, MD: US Dept of Health and Human Services;2007.

2. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA 2002guideline update for the management of patients with unstableangina and non–ST-segment elevation myocardialinfarction—summary article: a report of the American College ofCardiology/American Heart Association task force on practiceguidelines (Committee on the Management of Patients WithUnstable Angina). J Am Coll Cardiol. 2002;40:1366-1374.

3. Peterson ED, Roe MT, Mulgund J, et al. Association betweenhospital performance and outcomes among patients with acutecoronary syndromes. JAMA. 2006;295:1912-1920.

4. Katz DA, Dawson J, Beshansky JR, et al. Does concordance withguideline triage recommendations affect clinical care of patientswith possible acute coronary syndrome? Med Decis Making.2007;27:423-437.

5. Accreditation Process for the Society of Chest Pain Centers WebSite. Available at: http://www.scpcp.org/Accreditation/ChestPainCenters/tabid/60/Default.aspx. Accessed February 20,2009.

6. Hoekstra JW, Pollack CV, Roe MT, et al. Improving the care ofpatients with non–ST-elevation acute coronary syndromes in theemergency department: the CRUSADE initiative. Acad Emerg Med.2002;9:1146-1155.

7. Mehta RH, Roe MT, Chen AY, et al. Recent trends in the care of

Nonaccredited Centers(N�30,179)

Accredited Centers(N�3,059)

82.5 90.046.6 56.165.0 73.8

51.5 59.931.0 38.841.5 49.912.4 12.2

patients with non–ST-segment elevation acute coronary

Volume , . : July

Chandra et al SCPC Accreditation and NSTEMI Guidelines Adherence

syndromes: insights from the CRUSADE Initiative. Arch InternMed. 2006;166:2027-2034.

8. Liang KY, Zeger SL. Longitudinal data analysis using generalizedlinear models. Biometrika. 1986;73:13-22.

9. Armstrong PW, Fu Y, Chang WC, et al. Acute coronary syndromesin the GUSTO-IIb trial: prognostic insights and impact of recurrentischemia. Circulation. 1998;98:1860-1868.

10. Kwong RY, Schussheim AE, Rekhraj S, et al. Detecting acutecoronary syndrome in the emergency department with cardiacmagnetic resonance imaging. Circulation. 2003;107:531-537.

11. McCord J, Nowak RM, Hudson MP, et al. The prognosticsignificance of serial myoglobin, troponin I, and creatine kinase-MB measurements in patients evaluated in the emergencydepartment for acute coronary syndrome. Ann Emerg Med. 2003;42:343-350.

12. Fesmire FM, Hughes AD, Fody EP, et al. The Erlanger chestpain evaluation protocol: a one-year experience with serial 12-

lead ECG monitoring, two-hour delta serum marker

Volume , . : July

measurements, and selective nuclear stress testing to identifyand exclude acute coronary syndromes. Ann Emerg Med.2002;40:584-594.

13. Gibler WB, Blomkalns AL, Collins SP. Evaluation of chest painand heart failure in the emergency department: impact ofmultimarker strategies and B-type natriuretic peptide. RevCardiovasc Med. 2003;4:S47-S55.

14. Blomkalns AL, Gibler WB. Chest pain unit concept: rationale anddiagnostic strategies. Cardiol Clin. 2005;23:411-421.

15. Chen YP, Peto R, Collins R, et al. COMMIT (ClOpidogrel andMetoprolol in Myocardial Infarction Trial) collaborative group. Earlyintravenous then oral metoprolol in 45,852 patients with acutemyocardial infarction: randomised placebo-controlled trial. Lancet.2005;366:1622-1632.

16. Ross M, Lesikar S, Peacock WF, et al. Chest Pain Centeraccreditation is associated with better performance of Center forMedicare and Medicaid Services core measures for acute

myocardial infarction. Acad Emerg Med. 2007;14(suppl 1):55-56.

Annals of Emergency Medicine 25