Using Quality Improvement Methods to Improve Door-to-Balloon
Time at an Academic Medical Center

An estimated 500,000 patients per year in the United States¹ experience a ST-elevation MI (STEMI) with a shortterm mortality rate of 5–10%.² Percutaneous coronary intervention (PCI) is the preferred approach for treating STEMI¹ and timely reperfusion in patients with STEMI has been shown to increase the likelihood of survival.³ Time-toreperfusion is commonly measured by door-to-balloon time, which measures the time between a patient’s arrival at the hospital to when a balloon is inflated in the coronary artery. Door-to-balloon time for STEMI is a national quality measure reported to both the Centers for Medicare and Medicaid Services (CMS) and the Joint Commission on Accreditation of Hospital Organizations (JCAHO). The American College of Cardiology/American Heart Association (ACC/AHA) have recognized and recommended that doorto- balloon time should be less than 90 minutes (min).¹ The majority of hospitals do not meet the national guidelines^4–6 and recent progress has not been promising.⁶ Previous studies have delineated strategies to improve door-to-balloon times.^7–12 However, implementing these strategies remains a significant challenge.
A well-established approach to successfully implementing evidence-based strategies incorporates quality improvement methods using flow charts, plan-do-study-act (PDSA) cycles, measurement of processes and outcomes, implementation of change ideas, and rapid-cycle feedback.¹³ The Door-to-Balloon Alliance, an American College of Cardiology (ACC)-sponsored initiative, is another national effort to systematically improve door-to-balloon times using methods from quality improvement.¹⁴ In this study, we describe our experience implementing evidence-based strategies at a metropolitan academic medical center using this approach to decrease door-to-balloon time and measure improvement.

Methods
We used Vanderbilt University Medical Center’s (VUMC) cardiac procedure registry to identify the 56 consecutive patients who presented with STEMI from June 1, 2005 to November 30, 2006. The registry contains door-toballoon times and is the reporting source for JCAHO and CMS. Patients were eligible if: 1) they presented to VUMC with a STEMI (diagnosed using clinical presentation and initial electrocardiogram [ECG] with at least 1 mm ST-elevation in 2 contiguous leads, or new left bundle-branch block [LBBB]); 2) had symptom onset to presentation £ 24 hours; and 3) underwent subsequent cardiac catheterization with intervention. All STEMI patients were treated by primary PCI, and thrombolytics were not used for these patients at our institution. ECGs used were closest to presentation and none of the patients presented with LBBB or posterior MI. Exclusion criteria included transfer from an outside hospital (n = 4), ST-elevation after initial ECG (n = 3), acute MI symptom onset after arrival to the emergency department (ED) (n = 3), no PCI performed (n = 1), and delay in decision about consent (n = 1). The final cohort consisted of 44 patients over the 18-month period.
Quality improvement steps. Since the majority of our door-to-balloon times were not within the ACC/AHA recommendation of 90 min, administrative leadership charged both the cardiovascular medicine and ED to reduce door-to-balloon times. A quality improvement team was assembled and composed of the key stakeholders including the cardiac catheterization laboratory (CCL) director, ED director, cardiovascular medicine director, head cardiovascular case manager, cardiovascular medicine fellowship director and chief cardiac fellow.

A formal process analysis (Figure 1) was conducted with collection of baseline data on subinterval time periods and total door-to-balloon times from June 2005 to January 2006. Prior work on subinterval time periods¹⁵ identified achievable goals for each of the interval periods: 7.8 min for door-to-ECG, 45.3 min for ECG-to-CCL, and 29.0 min for CCL-to-balloon time, which were used to identify areas for improvement and set goals. In February 2006 systemwide changes to our former process of treating patients with STEMI were implemented (Figure 2). Based on previous work done by Bradley and colleagues,¹⁶ our process flow diagram and baseline data collection, we decided that door-to-balloon times could be improved by implementing the following: 1) activating the CCL by an ED attending physician; 2) insisting that a door-to-ECG time be less than 10 min (time when a patient first walks through the ED door to when he/she gets an ECG); 3) allowing 2 staff members tomove the patient from the ED to the CCL; 4) insisting on a 30- minute arrival time for CCL personnel including the interventional attending physician; 5) empowering the cardiology fellow to stop the process once it is started; 6) allowing the interventional cardiac fellow to obtain vascular access as soon as the interventional attending physician is present (not scrubbed); 7) using a diagnostic catheter for nonculprit arteries and an interventional guide catheter for the suspected culprit artery (prior to the intervention, the choice of catheter was operator-dependent); 8) providing real-time data feedback in a public forum (the monthly morbidity and mortality [M&M] conference); and 9) reviewing all cases with door-to-balloon times > 90 min.
When a STEMI patient was identified, ED activation of the CCL by the ED attending physician involved an operator delivering a simultaneous page to all members of the CCL team: cardiac fellow, interventional cardiac fellow, interventional attending physician, medical technicians and nurses. While the CCL was being activated, the patient would undergo a standardized workup and treatment plan including labs, chest X-ray, informed consent, and appropriate medications by both the ED and cardiology fellow to ensure that nothing would be missed or contraindicate the patient for cardiac catheterization. The cardiology fellow was able to stop the process if the patient was unsuitable for catheterization.
After the changes occurred, real-time data feedback was given to the CCL director and ED director. This data was shared with the entire cardiovascular medicine department during the monthly M&M conference. All cases that went over 90 min were reviewed by the CCL and ED directors within a week to identify problems or barriers in the process. In addition, comments and suggestions were collected for each patient undergoing the new process to achieve rapid-cycle improvement in the door-to-balloon process. From these comments, minor changes to personnel roles in the ED were made (i.e., who prepped the groin, who moved the patient, etc.). Statistical process control (SPC), which has been described elsewhere^17–19 and is commonly used in industrial engineering to measure process steps, was used to monitor overall door-to-balloon times in real-time.
Measures. The primary outcome measure was overall door-to-balloon time. Baseline preintervention times were recorded from the computerized medical record. A form was used to capture times throughout the door-to-balloon process postintervention. Subinterval time periods (door-to-ECG time, ECG-to-CCL time and CCL-to-balloon time) were measured to show what effect the changes had on the sub-processes in the overall door-to-balloon process. To evaluate the success of our interventions, dummy variables were created to compare the percentage of subinterval time periods at the goal defined by the prior process analysis. Secondary outcomes (death, coronary artery bypass grafting [CABG], intra-aortic balloon pump [IABP] placement, length of stay and inappropriate CCL activations) were collected to detect unintended consequences of the changes; however, our study was not powered to show differences in secondary outcomes. Baseline patient characteristics and presentation symptoms were collected to ensure that patients before and after the change were equal.
Sample size calculation. To detect a reduction in the mean door-to-balloon time of 25 min from the baseline 115 min with a standard deviation of 41 min under the assumption of 2 independent samples of equal sizes with a power of 0.80 and a = 0.05, we calculated a sample size of 43 STEMI patients. We used Dupont and Plummer’s PS (Power Size and Sample Calculation) software to make this calculation prior to analyzing the study. We averaged between 2–5 STEMI patients per month, and thus anticipated an 18- month study period to achieve the desired sample size.
Statistical analyses. For overall door-to-balloon and subinterval time periods, we compared medians with interquartile (IQR) ranges preintervention and postintervention using Wilcoxon rank-sum tests. The use of median times and IQR were used to both reduce the undue influence of large outlier times and estimate the variation around the average. Dichotomous baseline patient characteristics, symptoms at presentation, and secondary outcomes were analyzed using Fisher’s exact test. All analyses were performed using Stata 9.0 SE (Stata, College Station, Texas).
Statistical process control (SPC) charts for overall door-toballoon time were generated and graphically displayed to monitor in real-time the effect of changes made to the current process. SPC has rules which define statistical significance that have been explained elsewhere.^17–19 Upper confidence limits (UCL) were computed three standard deviations from the mean based on SPC statistical rules. SPC charts were created with Microsoft Excel (Microsoft, Inc.) and SPC XL (Digital Computations. Inc.). SPC charts were mainly shown to our quality improvement committee, and were also displayed monthly at our chest pain center meeting and intermittently at morbidity and mortality conferences.
SPC is a methodology for monitoring a process to identify normal and special cause variation. Originally developed by Walter Shewhart in the 1930s to help AT&T make better telephones, SPC was developed to help distinguish true change (termed “special cause” variation) from random process noise (“normal cause variation”). When applied to any process, SPC uses the distribution of the variables to determine confidence limits and a mean value, which define the stability of the process. Then, set rules (such as 6 valuesbelow the mean line) define statistically significant changes in the process. The statistics of determining confidence limits and the rules of SPC goes beyond the scope of this paper; however, many healthcare organizations have been using SPC to improve their processes for years. In fact, the quality improvement programs such as Six Sigma use this methodology to monitor processes for improvement.

Results

Baseline patient characteristics and presentation characteristics are presented in Table 1. No differences in baseline and presentation characteristics reached statistical significance including symptoms of heart failure, infarct characteristics and left ventricular ejection fraction at presentation. Specifically, patients presenting on the weekend (29% vs. 33%; p = 1.00, respectively) and at off-hours/night-time (29% vs. 22%; p = 0.521, respectively) were comparable in both preintervention and postintervention groups.
Subinterval time periods preintervention and postintervention to the door-to-balloon processes are presented in Table 2. Preintervention door-to-balloon times had a median of 108 min (interquartile range [IQR] = 94–122 min), higher than the ACC/AHA recommendation of 90 min. After implementing changes in the door-to-balloon process as outlined, our median postintervention door-to-balloon time decreased by 44 min (p < 0.001) to 64 min (IQR = 56 to 94 min). In addition, postintervention subinterval time periods for ED-to-ECG time decreased by 7 min (p = 0.006) from 14 min (IQR = 8 to 21 min) to 7 min (IQR = 3 to 10 min); postintervention ECG-to-CCL time decreased by 18 min (p = 0.01) from 57 min (IQR = 40 to 74 min) to 39 min (IQR = 23 to 49 min); whereas postintervention CCL-toballoon time did not change significantly.
In addition to comparing median reperfusion times preand postintervention, we determined the frequency with which patients received care within guideline recommendations (Table 2). This frequency is a measure of reliability of our new system to deliver guideline-concordant care for these patients. The postintervention percentage of STEMI patients within ACC/AHA guidelines of door-to-balloon times £ 90 min increased from 24% to 74% (p < 0.001). For the subprocesses, ED-to-ECG time £ 10 min increased from 35% to 78% (p = 0.01); ECG-to-CCL time £ 45 min trended toward an increase from 35% to 67% (p = 0.06); whereas CCL-to-balloon time did not change significantly.

The SPC charts (Figure 3) graph individual door-to-balloon times and the moving range (the difference in times from one patient to the next) in the 44 STEMI patients. With changes to the old process, a split in the control chart was created with patient number 18, the first STEMI patient to undergo the new process. The SPC chart revealed a decrease in mean door-to-balloon time of 39 min (from 114 to 75 min). The upper confidence limit (UCL) decreased from 248 min to 143 min, demonstrating that we had improved performance by decreasing variation in our process. The moving range chart on the bottom of the chart demonstrates decreased variation as the UCL decreased from 164 min to 84 min and the mean difference went from 50 min to 26 min between individual patients. The 6 red dots after the interventions were implemented represent higher performance and most likely were due to repetition and reinforcing education of the ED staff to activate the CCL (identified through our monthly chest pain meeting). We did not have enough points to meaningfully split the confidence limits again, however.
There was one outlying time point in the preintervention group that was reviewed and did not have a good reason for exclusion. Statistical analyses were run without the point and the effects postintervention were still statistically significant. The data presented here includes this point to be representative of the results reported to the JCAHO and CMS.
There were no statistically significant differences in the proportion of preintervention and postintervention patients who underwent CABG (6% vs. 4%; p = 1.00), had IABP placement (6% vs. 7%; p = 1.00), or died (0% vs. 7%; p = 0.515). The two deaths in the postintervention group were reviewed — both patients had arrested outside the hospital prior to admission, presented in shock, survived the catheterization and intervention, and their cause of death was most likely the extent of their infarct and not related to changes in the door-to-balloon process. The difference in length of stay (3.0 days vs. 3.5 days; p = 0.114) between the two groups also was not statistically significant. The directors of the CCL and ED reviewed all the cases and did not identify any inappropriate activation of the CCL, judged by the clinical context and the ECGs in the STEMI patients over this time period.
The postintervention review of cases (n = 6) of door-toballoon times > 90 min revealed the following reasons for delay: 1) the ED did not activate the CCL; 2) need for intubation and stabilization of the patient prior to getting the ECG; 3) difficult catheterization/intervention due to prior interventions; and 4) difficult weekend and late-night activations of the CCL. Subinterval time period analysis confirmed the reason for delay within the specified time period for each case > 90 min.

Discussion
Although evidence-based strategies for decreasing door-toballoon times (i.e., CCL activation by the ED, having a formal process for activating the CCL, expecting £ 30 min arrival for CCL staff after activation, etc.) have been described;10 the actual implementation of these strategies at single institutions has not been described to date. Moreover, implementation of all these strategies to improve door-to-balloon times may not be applicable for every hospital. Therefore, we chose to implement some of these strategies and continually evaluate the changes using quality improvement methods of process flow-charting, subinterval process time measurement, rapidcycle feedback and real-time monitoring with SPC charts.Using this approach, we decreased overall median door-toballoon times by 44 min and increased the percentage of STEMI patients receiving PCI within 90 min from 24% to 74%.
Our study links changes in the process to the subinterval components of door-to-balloon time. By insisting on a doorto- ECG time of less than 10 min and providing real-time feedback to the ED, we saw a 7-min improvement in that subinterval time period. ED activation of the CCL, allowing 2 staff members to move the patient from the ED to the CCL, and insistence of CCL staff arrival within 30 min resulted in an 18-min improvement for the ECG-CCL subinterval. Allowing the interventional fellow to obtain vascular access and using an interventional catheter for suspected culprit lesions did not significantly improve, nor did it worsen the CCL-to-balloon subinterval. Additionally, real-time monitoring, rapid-cycle feedback and SPC charting most likely contributed to improvement in overall door-to-balloon time by keeping participants engaged and accountable.
Key quality improvement action steps included: 1) leadership buy-in; 2) engagement of all the key stakeholders; 3) baseline flow-charting and meticulous measurement of the total process and subprocesses; 4) a team-building approach and collaboration between the ED, cardiovascular medicine department, and CCL; 5) rapid-cycle implementation of feedback to improve processes; 6) real-time monitoring of the door-to-balloon time including case-by-case review of defects; and 7) commitment by leadership to make door-to-balloon time a public priority.
Leadership buy-in was important in identifying stakeholders in key areas (namely the ED, CCL and cardiovascular medicine department) so that initial changes could be made rapidly. By building a quality improvement team with key members from the different departments, we were able to ensure adequate representation and communication when changes to the process were made. Results from the realtime SPC charting and case-reporting of defects were discussed at our monthly chest pain center meeting that allowed us to identify changes that were successful as well as areas requiring improvement.
From these results, PDSA cycles were used to quickly identify problems, implement solutions and measure results. Initially, we found that the ED was not always activating the CCL, which led to increased education for the ED staff. The roles of the ED and CCL personnel were defined and made clear, as this confusion led to delay. Another issue was that the times were not standardized, thus we aligned our clocks to standardize time measurements. We also found that pages were being delivered sequentially by the operator, rather than in a burst fashion, thus we implemented a system so that all members of the STEMI team could be notified together. All these problems were identified in our meetings and addressed later in the process.
Contextual factors in this study are important for generalizability. VUMC is a large academic tertiary care referral center for Middle Tennessee; however, there are 2 other tertiary care centers in the area. The number of STEMI patients presenting to our ED varied during the study period and ranged from 0 to 9 patients per month. Other institutions (especially with lower volumes) may find that “hardwiring” system changes into current processes could be more difficult. Another particular challenge for our medical center is that every month, different attending physicians, fellows and residents from different departments are involved in the process. Educating various stakeholders about the changes was difficult and required frequent, consistent and solid communication. In response to the initial cultural barrier about the ED causing inappropriate CCL activations, our CCL director did not feel there was any inappropriate activation over the study period. In our particular context, we believe that we systematically redesigned our door-to-balloon process so that delays could be identified and reduced and ultimately lead to a more reliable system.
As an academic teaching institution, preserving the educational “value” of caring for STEMI patients for ED residents, internal medicine residents and cardiology fellows is a priority. The ED residents conferred with ED attending physicians prior to activating the CCL. Although having the ED residents activate the CCL could save time, we felt that having an ED attending physician supervise the decision could lessen unnecessary activation of the CCL and decrease the chance that a STEMI would be missed. Our cardiology fellows were also allowed to make the critical decision to stop the catheterization process, which could have led to a potential delay; however, losing this experience would diminish the educational opportunity for our trainees and be inconsistent with our academic mission. An attending interventional cardiologist would review the case after he/she had arrived to ensure that the fellow’s decision was appropriate.
Study limitations. Limitations include that this study involves only one center and has a relatively small number of patients. Recognizing that this is a single-center study, relevant contextual variables have been given to help reproduce these changes. Another limitation is the “pre-post” study design with its inherent biases regarding regression to the mean and secular trend. We used SPC charting to address secular trends and make statistical conclusions about whether the new changes or redesigned processes were effective. Over the study period, the cardiology fellow and interventional cardiologist did not stop the process despite reviewing all the cases. Our study also did not capture how often the ED failed to catch a STEMI. Since we do not have any inappropriate activations of the CCL or information on how often the ED missed a STEMI, it is possible that our sensitivity for activating the CCL was too low, thus some STEMI patients might have been missed.
Our results suggest that most of the improvement opportunities for door-to-balloon time rest in the door-to-ECG and ECG-to-CCL subintervals. Door-to-balloon time by our current process seems to approach an asymptote of 60 minutes.Another opportunity for improvement could be the use of ECGs from emergency medical services (EMS) to activate the CCL.¹⁰ We did not employ this strategy since all EMS in Nashville are not equipped with 12-lead ECG machines; however, we have established relationships with EMS and are examining how to transmit ECGs with good quality and the feasibility of equipping all EMS with 12-lead ECG machines. We are developing protocols to activate the CCL when STEMI patients are identified by EMS.
The landmark book Crossing the Quality Chasm²⁰ calls for redesigning healthcare systems to reliably deliver evidencebased medicine to every patient. National efforts sponsored by the ACC such as the Door-to-Balloon Alliance are calling for institutions to implement evidence-based changes and use quality improvement methods to reduce door-toballoon times. In our study, the changes made to the EDto- ECG and ECG-to-CCL subintervals were effective, leading to a decrease in the overall median door-toballoon time. In addition, variation in the door-toballoon process was decreased and led to an increase in reliability for delivering a door-to-balloon time £ 90 min for 74% of STEMI patients with the new process. Although various interventions can be made to improve door-to-balloon time, quality improvement methods add to these interventions by measuring changes and facilitating process improvement. By using these methods to implement evidence-based practices, institutions can create effective teams and systems that will not only increase the number of STEMI patients whose door-to-balloon time will be £ 90 min, but begin to deliver higher performance in other areas of quality.
Acknowledgements. Robert Huang was supported by the VA Quality Scholars, Tennessee Valley Healthcare Services during the preparation of this manuscript. Christianne Roumie is supported by VA Career Development Award 04-342-2. The VUMC cardiac catheterization laboratory staff, ED staff, cardiovascular medicine staff and fellows were instrumental in the success of implementing the changes outlined in this study.

References

1. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction--executive summary: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1999 Guidelines for the Management of Patients with Acute Myocardial Infarction). Circulation 2004;110:588–636.
2. Waters RE 2nd, Singh KP, Roe MT, et al. Rationale and strategies for implementing community-based transfer protocols for primary percutaneous coronary intervention for acute ST-segment elevation myocardial infarction. J Am Coll Cardiol 2004;43:2153–2159.
3. McNamara RL, Wang Y, Herrin J, et al. Effect of door-to-balloon time on mortality in patients with ST-segment elevation myocardial infarction. J Am Coll Cardiol 2006;47:2180–2186.
4. Burwen DR, Galusha DH, Lewis JM, et al. National and state trends in quality of care for acute myocardial infarction between 1994–1995 and 1998–1999: The medicare health care quality improvement program. Arch Intern Med 2003;163:1430–1439.
5. Jencks SF, Huff ED, Cuerdon T. Change in the quality of care delivered to Medicare beneficiaries, 1998–1999 to 2000–2001. JAMA 2003;289:305–312.
6. Rogers WJ, Canto JG, Barron HV, et al. Treatment and outcome of myocardial infarction in hospitals with and without invasive capability. Investigators in the National Registry of Myocardial Infarction. J Am Coll Cardiol 2000;35:371–379.
7. Bradley EH, Curry LA, Webster TR, et al. Achieving rapid door-to-balloon times: How top hospitals improve complex clinical systems. Circulation 2006;113:1079–1085.
8. Bradley EH, Herrin J, Mattera JA, et al. Hospital-level performance improvement: Beta-blocker use after acute myocardial infarction. Med Care 2004;42:591–599.
9. McNamara RL, Herrin J, Bradley EH, et al. Hospital improvement in time to reperfusion in patients with acute myocardial infarction, 1999 to 2002. J Am Coll Cardiol 2006;47:45–51.
10. Bradley EH, Herrin J, Wang Y, et al. Strategies for reducing the door-to-balloon time in acute myocardial infarction. N Engl J Med 2006;355:2308–2320.
11. Bradley EH, Nallamothu BK, Curtis JP, et al. Summary of evidence regarding hospital strategies to reduce door-to-balloon times for patients with ST-segment elevation myocardial infarction undergoing primary percutaneous coronary intervention. Crit Pathw Cardiol 2007;6:91–97.
12. Nallamothu BK, Bradley EH, Krumholz HM. Time to treatment in primary percutaneous coronary intervention. N Engl J Med 2007;357:1631–1638.
13. IHI. Testing Changes. 2006.
14. ACC. D2B: An Alliance for Quality. 2007.
15. Bradley EH, Herrin J, Wang Y, et al. Door-to-drug and door-to-balloon times: Where can we improve? Time to reperfusion therapy in patients with ST-segment elevation myocardial infarction (STEMI). Am Heart J 2006;151:1281–1287.
16. Bradley EH, Roumanis SA, Radford MJ, et al. Achieving door-to-balloon times that meet quality guidelines: How do successful hospitals do it? J Am Coll Cardiol 2005;46:1236–1241.
17. Benneyan JC, Lloyd RC, Plsek PE. Statistical process control as a tool for research and healthcare improvement. Qual Saf Health Care 2003;12:458–464.
18. Carey RG (ed). Improving Healthcare with Control Charts: Basic and Advanced SPC Methods and Case Studies. Milwaukee: American Society for Quality, 2003.
19. Shahian DM, Williamson WA, Svensson LG, et al. Applications of statistical quality control to cardiac surgery. Ann Thorac Surg 1996;62:1351–1358; Discussion 1358–1359.
20. IOM (ed). Crossing the Quality Chasm. 5th ed. Washington, D.C.: National Academy Press, 2001.