Measuring the Quality of Primary PCI for ST-Segment Elevation Myocardial Infarction: Title and subTitle BreakTime for Balance

Reperfusion therapy with percutaneous coronary intervention (PCI) for ST-segment elevation myocardial infarction (STEMI) has understandably received considerable attention in quality measurement and improvement initiatives. STEMI is a common condition with high mortality. The benefits of reperfusion therapy can be substantial but are time sensitive.¹ However, important gaps exist between the patterns of treatment in the community and the ideal benchmarks established in evidence-based practice guidelines.² The timeliness of primary PCI, or “door-to-balloon” time, has emerged as one of a handful of quality measures reported to the public for the care of patients with acute myocardial infarction³ and serves as the focus of both regional^{4 - 5} and national initiatives^{6 - 7} to improve care quality.

Although understanding and improving the time to primary PCI in patients with STEMI has been widely accepted as important, this single measure overlooks other critical aspects of the performance of complex systems providing this important treatment. For example, while door-to-balloon time is relevant for patients who receive PCI, it provides no insights into the upstream processes of patient selection. Specifically, the publicly reported measure does not reflect how often systems fail to provide PCI for patients who should be treated (“false negatives”) or refer patients for angiography when acute reperfusion is not necessary (“false positives”). Undivided attention to improving the speed of the process of performing PCI could result in adverse consequences for patient selection by reducing the focus on false-negative decisions and increasing the numbers of false-positive referrals. In failing to provide PCI to patients who could benefit, and by exposing others to the risks of unnecessary invasive procedures, poor patient selection could meaningfully erode the net benefits to patients and increase the costs of efforts to improve the quality of reperfusion therapy.

While false negatives have been recognized as a quality problem,^{8 - 9} little is known about the frequency of false positives in systems that provide primary PCI for STEMI. In this issue of JAMA, the study by Larson and colleagues,¹⁰ which reports in detail the frequency and characteristics of false-positive referrals for primary PCI in a regional system, is both important and timely. This study reviewed the records of patients with suspected STEMI referred to a single primary PCI center from 30 regional hospitals. Catheterization laboratory activation was performed by emergency department physicians providing care at the transferring hospital with a single telephone call, thus substantially reducing barriers to referral. The capacity of this regional system to provide timely PCI has been reported previously⁴ ; the current study provides a broader perspective on the program.

Among the 1335 patients referred over 3½ years, 14% had no clear culprit coronary artery lesion, 9.5% had no significant epicardial coronary artery disease, and 11.2% had negative cardiac biomarkers. Thus, regardless of the definition used, no fewer than 1 in 11 referrals were considered false positives. These findings contrast with the limited data suggesting that false alarms are relatively rare in systems that provide primary PCI.^{11 - 12} However, in these prior studies, false-positive laboratory activations were not the primary focus and ascertainment either did not include systematic surveillance of all referred cases or involved a relatively small number of cases.

The false-positive cases described in the study by Larson et al constitute a heterogeneous patient population. In some patients, such as those for whom the electrocardiogram was apparently misinterpreted, the potential benefits of angiography were not likely to outweigh the risks. In other patients, however, such as those ultimately diagnosed with stress cardiomyopathy or coronary artery spasm, coronary angiography was likely an appropriate diagnostic test even though PCI was ultimately not performed. Lacking hindsight, while some referrals designated as false positives were likely justifiable, it is clear that false positives constitute more than a trivial proportion of the population.

This finding supports the authors' suggestion that false-positive catheterization laboratory activations merit further scrutiny, complementing the time to primary PCI as a reflection of quality. Measures to quantify the undesirable consequences of quality improvement efforts—“balancing measures”—have been proposed in other contexts.¹³ Moving forward, such measures merit greater consideration with respect to both established and prospective areas of performance assessment. When there is evidence for substantial misuse of emerging therapies,^{14 - 15} the development of quality measures intended to encourage adoption could have foreseeable negative consequences unless simultaneous efforts to discourage misuse are also in place.

While collecting data on false-positive referrals for PCI would provide a broader perspective on quality of care, the appropriateness of public reporting of this measure is less clear. To some extent, “false-positive” catheterization laboratory activations are unavoidable. Like any diagnostic test, the decision to send a patient for urgent angiography based on suspicion of STEMI will be imperfect. Because of the necessary trade-offs between specificity and sensitivity, an absence of false-positive activations would imply the failure to provide PCI to some deserving patients. Even if complemented with an understanding of the failure to provide PCI to eligible patients, reporting false-positive referrals for primary PCI would likely be difficult to interpret, risking confusion over clarity. However, expectations that systems providing PCI perform surveillance for errors in patient selection would encourage an understanding of the reasons for such events and stimulate efforts to limit their occurrence.

The focus on processes of care as measures of the quality of primary PCI implicitly assumes that such process measures are reasonable proxies for patient outcomes. However, just as surrogates are considered poor substitutes for measuring the effect of drug therapy on meaningful clinical outcomes, the relationship between process measures and such outcomes has appropriately been subjected to greater scrutiny.^{16 - 17} Balancing measures, by acknowledging the possibility that perturbations in complex systems intended to improve care may also have negative consequences, are a step forward. However, they can only measure repercussions that are foreseen. The law of unintended consequences dictates that even the most elaborate balancing measures will not capture all the important ramifications of efforts to improve the delivery of PCI to patients with STEMI. Ultimately, the proof of the effectiveness of quality improvement efforts resides in an understanding of their effect on outcomes and the costs of achieving these benefits.

In the past, measuring outcomes of care has been hampered by both methodological and logistical constraints. However, the methods of outcomes measurement have matured.¹⁸ Furthermore, emerging data sources such as national registries hold promise for overcoming many of the logistical problems by allowing access to large numbers of patients seen in diverse practice settings.¹⁹ Thus, the use of outcome measures to complement process measurement is becoming more plausible.²⁰

The study by Larson et al should stimulate further development of quality measurement and reporting efforts. In the case of primary PCI, the view of quality should extend beyond the time to treatment to include patient selection, and ultimately to outcomes. While efforts to improve the quality of care will—and should—continue, the science of improvement must advance so that clinicians can be increasingly confident that efforts to deliver the right care to the right patient at the right time are improving and extending patients' lives.