Evaluating and Improving the Quality of Care for Acute Myocardial Infarction: Title and subTitle BreakCan Regionalization Help?

Certificate of need programs were initiated 3 decades ago in the United States for the purpose of limiting expenditures for particularly costly interventional cardiovascular services and technologies.¹ Certificates of need are a kind of regionalization, a process by which specialized procedures like coronary artery bypass graft (CABG) surgery and percutaneous coronary intervention (PCI) are deliberately distributed in a presumably rational and efficient geographic context. However, certificate of need regulations have fallen out of favor in many states; they have been repealed by 19 states entirely and by 25 states for CABG surgery.¹ Little has been done in recent years to examine the impact of certificates of need on cost or on quality of care except for studies related to CABG surgery and PCI.^{2 - 3}

In this issue of JAMA, Popescu and colleagues⁴ examine the association between certificates of need and treatment of a medical condition, acute myocardial infarction (AMI), that is not controlled by certificates of need but frequently involves the use of procedures controlled by certificates of need, namely, CABG surgery and PCI (collectively called coronary revascularization procedures). The authors found that in states with certificates of need, Medicare patients with AMI were less likely to be admitted to hospitals with coronary revascularization services than were patients in states without certificates of need and were also less likely to undergo revascularization at the admitting hospital. Although patients in the states with certificates of need were more likely to undergo revascularization at a transfer hospital, the overall effect was that patients in states with certificates of need were less likely to undergo revascularization within the first 2 days (adjusted hazard ratio, 0.68, 95% confidence interval, 0.54-0.87; P<.001 for patients in states with stringent certificate of need regulations relative to states without certificates of need). However, despite the lower rates of early revascularization, the risk-adjusted 30-day mortality was no different for patients in states with and without certificates of need (odds ratio, 1.00; 95% confidence interval, 0.97-1.03; P = .90).

As the authors note, 2 possible explanations for these seemingly contradictory findings are that (1) hospitals in states with certificates of need perform higher volumes of revascularization procedures, and higher volumes of these procedures are associated with better outcomes^{1 - 2} ; and that (2) some of the additional revascularization procedures performed in states without certificates of need are among patients who derive, at best, marginal benefit.

With regard to choice of nonbeneficial intervention, the International Classification of Diseases, Ninth Revision, Clinical Modification codes used to report diagnoses in the Medicare database used by the authors and in many other administrative databases identify the site of the AMI. The fourth digit in this 5-digit code represents the site of the AMI and a single site (fourth digit of 7) is associated with a subendocardial (or nontransmural) infarction. As of 2006, coders are specifically instructed to report non–ST-elevation AMIs (NSTEMIs) as subendocardial infarctions.⁵ Thus, STEMIs are to be assigned to one of the other sites. Although this reporting might have been used by some coders in earlier years, the 2006 data are more reliable in this regard because of the new instructions. Consequently, one possibility for estimating the appropriateness of the intervention with new 2006 data is to further subdivide the 30-day mortality analyses (by subendocardial infarction or other site) in addition to whether patients received revascularization. This would provide a hint as to whether patients with subendocardial infarction receive a mortality benefit from revascularization and whether a higher percentage of patients in non–certificate of need states are undergoing unnecessary revascularization procedures.

Several additional extensions of these analyses may be helpful. For instance, subsequent analyses could attempt to identify the contribution of superior revascularization in certificate of need states by comparing 30-day mortality rates in states with and without certificates of need separately for patients who did and did not undergo revascularization. This would allow for an estimate of any decrease in mortality related to more successful revascularization procedures in states with certificates of need. Also, the relationship between revascularization volume and 30-day mortality could be tested for states with and those without certificates of need.

The study by Popescu et al⁴ is particularly timely in view of recent efforts to regionalize AMI care. There are well-established guidelines for the care of AMI patients in general⁶ and STEMI patients in particular.⁷ Because evidence suggests that STEMI patients can benefit most if they have early access to PCI,⁸ there is a need to explore how this can be achieved, including the possibility of designating hospitals as STEMI centers, equipping ambulances with capabilities to perform 12-lead electrocardiograms, instructing paramedics to transport patients with STEMI to STEMI centers, and decreasing door-to-balloon times for AMI patients undergoing primary PCI. Efforts sponsored by the American Heart Association are now under way to convene stakeholders to help develop optimal delivery systems for STEMI patients.⁹

Also, a recent study indicates that the success of prehospital triage protocols for patients with STEMI will partially depend on how patients are geographically distributed relative to hospitals that perform PCI; that study found that nearly 80% of the adult US population lived within 60 minutes of a PCI hospital in 2000.⁹ Furthermore, even among those living closer to non-PCI hospitals than to PCI hospitals, almost three fourths would have less than a 30-minute additional delay with direct referral to a PCI hospital.¹⁰

These 2 studies^{9 - 10} underscore the importance of the findings by Popescu et al⁴ because they suggest that criteria for further regionalizing care for AMI are needed and will probably be more prevalent in the future. This will require identification of the optimal number and location of STEMI centers so that patients can be transported to these centers quickly while maintaining high enough volumes to ensure low rates of adverse outcomes. If PCI hospitals are not well distributed geographically in some states with certificates of need, the population who lives within 60 minutes of a PCI hospital could be considerably lower than it is in states without certificates of need. This would argue against certificates of need or, at least, for the approval of more PCI/STEMI hospitals in states with certificates of need. Conversely, for states with certificates of need in which the geographical distribution of PCI is reasonably good, it could be easier to establish STEMI centers because most or all current PCI hospitals could be designated as STEMI centers, and ambulance triage would be the main part of the system that required further organization.

For states without certificates of need, there may be a problem determining the subgroup of PCI hospitals to designate as STEMI centers. Any tendency to designate too many STEMI centers based on political and other pressures could lead to worse outcomes resulting from low volumes and the well-known volume-outcome relationship for PCI in AMI patients.^{11 - 12}

To their credit, Popescu et al acknowledge the limitations associated with using administrative data to analyze clinical outcomes. These limitations should be considered seriously when planning research agendas for evaluating the effectiveness of regional systems and the effectiveness of individual hospitals in caring for AMI patients. Although it is generally agreed that chart-based registries are the standard approach for evaluating inpatient care, this option is prohibitively expensive without electronic medical records. Consequently, the use of administrative data for this purpose would appear to be the only feasible option. It would be more helpful to consider appending a minimal number of clinical data elements to administrative databases for purposes of evaluating the quality of AMI care, and this option is currently under consideration in New York State. Such data elements include systolic blood pressure, time from onset of symptoms to arrival at the hospital, and creatinine and blood urea nitrogen levels (although ST elevation is also an important predictor of AMI outcomes, it can now be obtained from administrative data).

It would also be valuable if a limited set of process data could be appended to administrative data sets. For example, it is well established that use of aspirin and β-blockers^{7 ,13} at the times of hospital admission and discharge are related to better outcomes, and it would be important to know if there are differences in the use of these processes of care in states with certificates of need and those without certificates of need. Time between arrival at the hospital and delivery of fibrinolytic therapy (door-to-needle time) and time between arrival and PCI (door-to-balloon time) are also important process measures.^{14 - 15} It would also be important to relate interhospital differences in the use of these processes to differences in short- and long-term outcomes for quality improvement purposes, and to determine if there are differences in the use of these processes in states with and without certificates of need.

In conclusion, regionalization is an important concept in the delivery of optimal care for patients with AMI. Certificates of need for revascularization procedures are a form of regionalization for AMI because these procedures are appropriate for some AMI patients. The study by Popescu et al demonstrates that government-based regionalization is safe for AMI patients in that 30-day mortality is not compromised despite lower revascularization rates. The next challenge is to determine how to fine-tune regionalization policies in certificate of need and non–certificate of need states to create systems that are as effective and efficient as possible.