Author + information
- Theodore A. Bass, MD∗ ()
- Department of Medicine, Division of Cardiology, University of Florida College of Medicine-Jacksonville, Jacksonville, Florida
- ↵∗Address for correspondence:
Dr. Theodore A. Bass, Department of Medicine, Division of Cardiology, University of Florida College of Medicine-Jacksonville, 655 West 8th Street, Jacksonville, Florida 32209.
Percutaneous coronary intervention (PCI) has continued to evolve throughout its history, resulting in vastly improved procedural success rates and clinical outcomes with lower procedural complications compared with results initially reported 4 decades ago. This improvement has continued to grow in a setting of treating an expanded, more complex, and higher risk population. Much of this success has been attributed to the robust innovative improvement in device technology, cardiovascular imaging, and adjunctive pharmacotherapy, in addition to our greater understanding of pathophysiology. All of these innovations come with a considerable financial price tag. We currently practice in an era of increased and often compulsory data reporting, public reporting of clinical performance measures, and increasing scrutiny from regulators, payers, and well-informed patients. Indeed, to correctly code and report these data requires additional resources, adding further cost to maintaining a PCI program in this complex environment.
The interventional cardiovascular community has spent considerable effort to better understand what factors contribute to attaining optimal clinical outcomes in patients presenting with coronary artery disease treated with PCI. These efforts have frequently involved the analysis of data obtained from large databases, many of which are administrative in nature, providing little clinical insight into patient or procedural risk. This has made it increasingly difficult to risk adjust results and evaluate procedural performance lacking adequate important clinical context. For example, asking the question “Is individual PCI operator volume or institutional hospital volume correlated with clinical procedural outcomes?” has yielded unclear and often conflicting signals (1,2). Is this because there is no correlation, or are we perhaps unable to adequately risk adjust the populations to determine whether, indeed, there is a volume relationship to outcome? The American College of Cardiology’s National Cardiovascular Data Registry (NCDR) currently offers our best available dataset to develop risk-adjusted models necessary to critically compare clinical performances. However, not surprisingly, there are still gaps in these datasets that limit our ability to adequately risk adjust when making comparisons dealing with the complexities of clinical care and outcome measurements.
In this issue of JACC: Cardiovascular Interventions, Acharya et al. (3) compare the risk-adjusted PCI outcomes of safety-net hospitals (SNH) and non-SNH. The investigators use the Institute of Medicine definition of a SNH, a hospital with at least 10% of their PCI volume involving uninsured patients. Using the NCDR Cath-PCI Registry from 2009 to 2015, they analyzed 3,746,961 patients who underwent PCI at either 282 SNH or 1,134 non-SNH, finding a “marginally” higher risk-adjusted in-hospital mortality at SNH (odds ratio: 1.23; 95% confidence interval: 1.17 to 1.32; p < 0.001) and a 0.4% higher absolute risk of in-hospital mortality. This finding was most marked comparing lower PCI-volume hospitals and appeared to be mitigated as PCI institutional volume increased. The investigators also reported that bleeding and acute kidney injury rates were similar, showing no significant difference comparing SNH to non-SNH. The authors acknowledge that any dataset used to develop a risk-adjustment model in this very complex SNH population will have limitations. They conclude that it is reassuring that SNH “despite treating a complex and acutely ill population are able to achieve satisfactory if not similar (PCI) outcomes as non-SNH.” Reporting the observed to expected outcomes, the investigators also suggest that the current NCDR risk model can serve as a foundation for future SNH and non-SNH comparisons.
The authors should be congratulated for bringing up an import topic: the quality of interventional cardiovascular health care at hospitals that serve an often overlooked population of unfunded or underinsured patients. These patients are particularly vulnerable, frequently having little or no ability to select their access to health care institutions or medical providers on the basis of quality outcomes. This report is particularly timely in light of recent and possible future health care legislation that can seriously affect the resources made available to these hospitals. Overall, the authors correctly acknowledge the inherent limitations of the study, mostly involving the issues of SNH definition in the current U.S. health care matrix and the challenges of risk adjusting the medically complex underfunded population often treated at these institutions. It is encouraging that despite a statistically significant difference due to the large number of cases, overall, there is only a 0.4% absolute difference in mortality. What can we learn from this analysis other than reassurance that patients presenting for PCI at this SNH group appear to receive similar safe care compared with patients treated at a non-SNH?
Transparency is necessary to put my comments in proper context when addressing perceived limitations of this report. For over 3 decades, I have spent my entire medical career as a practicing interventional cardiologist at an academic, teaching SNH located in an inner city in the southeastern United States. A significant part of my practice involves tertiary referrals for resource-demanding, high-risk complex interventional coronary procedures. For the past 25 years, I have also had administrative assignments involving oversight of the cardiovascular service line including quality and financial responsibilities.
Although the Institute of Medicine definition of a SNH used by the investigators does apply to my institution, it offers limited insight into the funding and resource challenges we may or may not face. There is no uniformly accepted definition of a SNH in the United States. The Safety Net Hospital Alliance of Florida (SNHAF) is an advocacy group representing the 14 SNH throughout the state that comprise <10% of the hospitals in Florida. These SNH provide over 50% of the charity care and nearly 50% of all Medicaid hospital care in the state. Seventy-five percent of their patients are government-sponsored or uninsured. Their Medicaid/uninsured patient caseload is 66% higher than the rest of the hospital industry in the state (4). However, this is a very diverse collection of hospitals, including teaching and non-teaching hospitals, publically funded institutions supported by local city and county tax revenues, and many that have no such support. Hospitals in this or any SNH group may have very different obligations and exposure caring for the large pool of unfunded or poorly funded patients throughout their region or state. Some of these institutions may have other public service commitments, such as regional high-end trauma centers that compete for resources in an already strained financial system. I suspect this heterogeneity is even more pronounced when we group SNH nationally. When investigating quality of care, there are several different, commonly used approaches used for classifying hospitals as SNHs. These models can involve multiple different identifying characteristics involving the hospital commitment and burden in providing uncompensated or poorly compensated care. These SNH model definitions may include such factors as a hospital’s relative contribution to the total amount of uncompensated care provided in the community, the comparative Medicaid caseload compared with statewide standards, or may be based on other characteristics such as being labeled as a public hospital or a member of the Council of Teaching Hospitals and Health Systems. Not surprisingly, quality of care findings can vary depending on which SNH definition is used (5). These funding and service characteristics are important to help appreciate the differing financial challenges faced by individual institutions and to better appreciate the importance of finance and resource availability in achieving desirable quality outcomes. Understanding the heterogeneity of the SNH group represented in this report is a challenge. It would be helpful to know the variance of clinical outcomes in the SNH group. Were there identifying characteristics other than low institutional volume of the poorly performing outliers within this group?
It is known that underinsured patients tend to be sicker and present with more and often-untreated comorbidities. This observation has been used to explain less favorable medical outcomes realized in this complicated population. When developing risk models dealing with these patients, the dichotomous recording of the presence of cardiovascular risk factors and comorbidities does not provide a complete evaluation of risk. The diabetic, hypertensive, and dyslipidemic patients comprising this underinsured population are often suboptimally treated for these conditions due to educational, financial, or social challenges. There are gradients of risk for hypertensive patients based on their blood pressure control. The obese are often morbidly obese. Patients having undergone a successful PCI at a SNH might more often succumb to other untreated conditions while hospitalized, such as hypertensive stroke or pneumonia. It is not unusual to receive patients from the trauma service experiencing acute myocardial infarction in conjunction with their major trauma incident. Many of these additional quantitative confounders are not entered into the NCDR database. How are these risks adjusted?
Finally, it is hard to overstate the importance of having adequate resources to properly code and enter data for appropriate risk adjustment in this very complex population. SNH tend to have smaller gains in quality of care performance measures than non-SNH (6). Many SNH do not have the necessary resources to develop and maintain the quantity and quality of support personnel, technical support, consulting help, and infrastructure required to compete with well-funded private institutions that are fully committed to these processes in the era of public reporting.
This study helps shed more light on the potential issues of health care disparity in cardiovascular interventional care in the United States. Although the lack of any strong signal suggesting inferior care is not a surprise, it is important to appreciate the great heterogeneity of SNH in the current health care system to better understand and best address quality of care issues. It may well be that, not only are PCI outcomes noninferior at SNH compared with non-SNH, the sicker, more complicated patient with more comorbidities might perhaps be better served at a SNH.
↵∗ Editorials published in JACC: Cardiovascular Interventions reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Interventions or the American College of Cardiology.
Dr. Bass has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation
- Harold J.G.,
- Bass T.A.,
- Bashore T.M.,
- et al.
- Badheka A.O.,
- Patel N.J.,
- Singh V.,
- et al.
- Acharya T.,
- Salisbury A.C.,
- Spertus J.A.,
- et al.
- Safety Net Hospital Alliance of Florida