Root Cause Analysis of Deaths in ST-Segment Elevation Myocardial Infarctions Treated with Primary PCI: What Can We Do Better?

Abstract: Objective. Recent data demonstrate that mortality of patients with ST-elevation myocardial infarction (STEMI) has not changed despite dramatic reduction in door-to-balloon times. Identifying potential areas in care that can be further optimized to decrease mortality remains a priority. Methods. We performed a root cause analysis of all patients who died following primary percutaneous coronary intervention (PCI) during index hospitalization from 2008 to 2013 at the University of Michigan. Using a standardized data collection form, two interventional cardiologists and one non-invasive cardiologist reviewed patient care prior to arrival to the catheterization lab, while in the catheterization lab, and after primary PCI to determine cause of death and to rate potential preventability of death on a Likert scale (0 unpreventable—4 mostly preventable). Results. Of the 25 deaths over the 5-year period, 8 were deemed at least mildly preventable by one or more reviewer. No death was deemed totally preventable. Interreviewer agreement was moderate for both cause of death (nominal Krippendorff’s alpha = .58) and preventability of death (nominal alpha = .233). In spite of this overall lack of agreement, in all 8 preventable cases at least one reviewer cited ischemia to balloon time as a potentially addressable factor associated with the death. Conclusion. Mortality following primary PCI was deemed mostly unpreventable. However, improvement in total ischemic time, and in particular symptom-onset to medical care, was identified as one potential target that might be of value in further reducing the mortality associated with STEMI.

J INVASIVE CARDIOL 2017;29(5):164-168. Epub 2017 March 15.

Key words: angioplasty, cost, primary PCI, mortality, STEMI

Timely reperfusion is the cornerstone for the treatment of ST-elevation myocardial infarction (STEMI).¹ In-hospital mortality secondary to STEMI decreased significantly in the 1990s and early 2000s,² largely due to widespread use of reperfusion therapy initially in the form of fibrinolytics and subsequently due to widespread adoption of primary percutaneous coronary intervention (PCI). As part of the emphasis on timely reperfusion, American College of Cardiology (ACC)/American Heart Association (AHA) guidelines gave a strong endorsement for door-to-balloon time of 90 minutes or less.¹ This became a national quality measure and a key target for health-care systems. However, recent data, both in a regional and a national review, show that despite 80% of primary PCIs performed on STEMI patients occurring within 90 minutes, mortality rates failed to improve.^3,4 It is unclear whether there is a quality gap that if bridged might help improve outcome of patients presenting with STEMI. The goal of our study was to perform a root cause analysis of all deaths occurring following primary PCI at a single academic center to determine possible preventability of death and to identify future targets for treatment or quality improvement. 

Methods

This study included all patients presenting to the University of Michigan with STEMI from January 2008 to July 2013 who died during index hospitalization following primary PCI. Data from the Blue Cross Blue Shield of Michigan Cardiovascular Consortium were used to identify these select patients. 

A data collection form was created (Appendix 1) and each patient’s care prior to hospital arrival, in the emergency department, prior to arrival to the catheterization lab, and in the catheterization lab was examined in detail. 

Three cardiologists (two interventional and one non-invasive) reviewed the data collection form, the discharge summary, and the catheterization report for all included patients. The reviewers evaluated each case to determine the cause of death (classified as cardiogenic shock, neurological, respiratory, infectious, hemorrhagic/vascular, or other) and determined if each death was preventable using a Likert scale (0 = unpreventable; 1 = slightly preventable; 2 = moderately preventable; 3 = mostly preventable; 4 = entirely/certainly preventable). If a case was deemed to be preventable, each phase of care (prehospital, emergency room, catheterization laboratory, post catheterization laboratory) was rated on the same Likert scale in likelihood of preventability of death. In cases that they defined as potentially preventable, reviewers were asked to further elaborate in text form the improvement in care with which survival would have been improved. 

Given the subjective nature of our review, interreviewer reliability was quantified using the average pairwise agreement percentage and the Krippendorff’s alpha coefficient, given their ability to evaluate nominal values.^5,6 These were calculated in order to evaluate for potential high variability in causes of death and assess consistency in deaths thought to be preventable. Given the small sample size, we considered all deaths that were rated as ≥2 by at least one reviewer as preventable.

Results

During the study period, there were 629 patients who underwent primary PCI, of whom 25 died during the index hospitalization (3.9 %). We divided patients into those whose death was deemed non-preventable and those whose death was potentially preventable. Baseline characteristics are listed in Table 1 and overall presentation times are noted in Table 2, with two groups presented (those deemed preventable and unpreventable). The clinical scenarios for all preventable cases are summarized in Table 3. The average age for the preventable group was 67.8 years, with 42.9% women. The average age for the unpreventable group was 68 years, with 29.4% women. The preventable group had a median ischemia-to-hospital time of 189 min, door-to-balloon time of 52 min, and total ischemia-to-balloon time of 311 min. The unpreventable group had a median ischemia-to-hospital time of 70 min, door-to-balloon time of 76.5 min, and ischemia-to-balloon time of 160 min. Death was rated as cardiogenic in etiology most frequently, ranging from 60%-72% between reviewers. For cause of death, the average pairwise percent agreement was 78.67% with no obvious difference between interventional and non-invasive cardiologists. The Krippendorff’s alpha coefficient was 0.58 among the three reviewers. 

A distribution rating of preventability of death by reviewer is presented in Figure 1. No case was deemed totally preventable by any reviewer and there were 8 deaths that were deemed at least moderately preventable by at least one reviewer. The average pairwise agreement among all three reviewers on preventability of death was 61.33% and Krippendorff’s alpha coefficient test for interreviewer variability was 0.233, with no obvious difference noted between interventional and non-invasive cardiologists. 

Decreasing symptom onset-to-medical contact time or ischemia-to-balloon time was rated most commonly as a modifiable factor, with at least one reviewer identifying it as potentially important in each of the 8 cases. Other potential areas of improvement were earlier recognition of hemorrhagic shock, earlier recognition of in-stent thrombosis, decreased time in the emergency department, and possibly the use of radial approach in 1 case complicated by a retroperitoneal hematoma. 

Discussion 

Short-term mortality following STEMI has declined to about 4.5%-5% in the era of widespread primary PCI.² Although this is a significant improvement over the past several decades, this has not improved significantly in recent years despite improved institutional compliance of door-to-balloon time <90 minutes and a decrease in overall average door-to-balloon times regionally and nationally.^3,4 This study evaluated each individual case of in-hospital death following STEMI treated with primary PCI to perform a root cause analysis to elucidate areas of care that may allow for improved outcomes. Our reviewers noted that a total of 8 out of the 25 patient deaths were partially preventable, as graded by at least one of the three reviewers. 

However, upon review of causes that could potentially improve outcomes, our reviewers showed a strong trend in that a decrease in symptom onset-to-angioplasty time could improve outcomes. For all 8 cases, at least one of the reviewers noted that a decrease in ischemia to hospital time would have likely improved outcomes. To a lesser extent, our reviewers identified 2 cases when earlier recognition of STEMI in the emergency department may have contributed to preventing death. This is consistent with our data, which showed patients who were thought to be preventable did have a higher ischemia-to-hospital time (189 min vs 70 min) and total ischemia-to-balloon time (366.3 min vs 260.7 min). 

Despite national averages of door-to-balloon times reaching 67 min in 2008-2009,³ these numbers do not include total ischemic time. These trends do raise the question as to whether total ischemic time should be a community performance measure, and if so, whether reducing it would improve mortality. 

Our data suggest that increased ischemia-to-hospital times and ischemia-to-balloon times are potentially improvable phases of care that may currently be contributing to worsened outcomes in patients presenting with STEMI and treated with PCI. Focus on improvements in patient recognition of symptoms before health-care contact, encouragement of proper transportation to medical facilities, and quicker diagnosis upon medical contact could help further drive down ischemia-to-balloon time. The realization that increased ischemia-to-hospital time leading to worse outcomes is not novel, and is identified by current management guidelines for STEMI management.¹ However, previous efforts to improve ischemia-to-hospital time by increasing community awareness, education on symptom recognition, and providing direction to initiate management have been performed in an era that is much different than the era we live in today.⁸ The ubiquity of smart phones, electronic tablets, and computers may allow for new avenues and greater influence to improve ischemia-to-hospital time and potentially improve outcomes following MI. 

Study limitations. Our study is not without limitations. This was a retrospective study of a single institution; hence, despite a 5-year span, our study population is limited and lacks power. Reviewers were not blinded to catheterization reports or colleagues who treated the patients. Given that this study is a root cause analysis, these limitations, although noted, are not felt to hinder the conclusion that further focus on a community level to improve symptom recognition in the general population and to further investigate how to improve ischemia-to-balloon time is warranted. 

References

1.    O’Gara PT, Kushner FG, Ascheim DD, et al. 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2013;61:e78-e140.

2.    Roe MT, Messenger JC, Weintraub WS, et al. Treatments, trends, and outcomes of acute myocardial infarction and percutaneous coronary intervention. J Am Coll Cardiol. 2010;56:254-263.

3.    Flynn A, Moscucci M, Share D, et al. Trends in door-to-balloon time and mortality in patients with ST-elevation myocardial infarction undergoing primary percutaneous coronary intervention. Arch Intern Med. 2010;170:1842-1849.

4.    Menees DS, Peterson ED, Wang Y, et al. Door-to-balloon time and mortality among patients undergoing primary PCI. N Engl J Med. 2013;369:901-909.

5.    Freelon D. ReCal: intercoder reliability calculation as a web service. Internat J Internet Sci. 2010;5:20-33.

6.    Freelon D. ReCal OIR: ordinal, interval, and ratio intercoder reliability as a web service. Internat J Internet Sci. 2013;8:10-16.

7.    Cannon CP, Gibson C, Lambrew CT, et al. Relationship of symptom-onset-to-balloon time and door-to-balloon time with mortality in patients undergoing angioplasty for acute myocardial infarction. JAMA. 2000;283:2941-2947.

8.    Simons-Morton DG, Goff DC, Osganian S, et al. Rapid early action for coronary treatment: rationale, design and baseline characteristics. Acad Emerg Med. 1998;5:726-738.

From the Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, Michigan.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted November 11, 2016; provisional acceptance given November 21, 2016; final version accepted November 26, 2016.

Address for correspondence: Fredy El Sakr, MD, 1500 E Medical Center Dr, Ann Arbor, MI 48109. Email: felsakr@med.umich.edu