CMS Penalties Spur Efforts to Reduce 30-Day Hospital Readmissions

With the advent of the Affordable Care Act and public reporting of readmission rates by the Centers for Medicare & Medicaid Services (CMS), healthcare professionals and researchers are looking more closely at how to prevent patients from unnecessarily returning to the hospital after their initial discharge.

In a 2-part series, TCTMD is exploring excess readmissions and what can be done about them.

Part 1: The Case for Readmission as a Quality Metric

An editorial published last year suggests that hospitals are likely to increasingly focus efforts to reduce readmission on PCI patients. Furthermore, the American College of Cardiology and CMS have collaborated to implement voluntary public reporting of 30-day readmission rates after PCI, the editorial reports.

Hospital readmission, estimated to occur in the United States at an incidence of 8%-16% within 30 days of PCI, has been linked to increased 1-year mortality in multiple studies. Moreover, post-PCI rehospitalization is expensive, accounting for a substantial portion of the $26 billion that Medicare pays annually for readmission of fee-for-service patients.

CMS assesses a penalty—currently up to 3% of all Medicare reimbursement—on hospitals whose readmission rates exceed the national average, based on a formula that sums up performance on 5 covered conditions. According to the American Hospital Association, today 78% of all hospitals are being hit with some level of penalty.

The diagnoses now under scrutiny are acute MI, heart failure, pneumonia, COPD, and hip and knee replacement. But the Affordable Care Act authorizes penalties to be extended to all conditions—including all those treated with PCI—and that expansion is likely to be the next step in the CMS campaign, several experts told TCTMD.

Noting that PCI has been recognized by the Medicare Payment Advisory Commission (MedPAC) as an important source of Medicare spending, Jason H. Wasfy, MD, MPhil, of Massachusetts General Hospital (Boston, MA), advocates its inclusion in the CMS program. “If you just focus on acute MI, you’re missing a lot of the PCI population, and there is a lot of value improvement to be had in those patients,” he told TCTMD in a telephone interview.

Policy Creates a ‘Period of Pain’

The CMS policy “is creating great financial pressure on hospitals to be creative in how they think about an entire episode of care rather than a simple cross section, which is how we normally think of hospitalization,” John A. Spertus, MD, MPH, of St. Luke’s Mid America Heart Institute (Kansas City, MO), said in a telephone interview with TCTMD.

“In theory, I like that [pressure] a lot. But in the short run, it creates some challenges,” he noted. “Hospitals are strongly incentivized to come up with solutions and invest in programs. But we haven’t yet figured out how to reduce readmissions very well. So it’s creating a period of pain for hospitals.”

In addition, overzealous efforts could potentially put patients at risk, he suggested. For example, an acute MI patient who returns to the hospital with heart failure may be held in the emergency department (ED) for several days to try to avoid an official readmission. “That’s masking the problem, not solving it,” Dr. Spertus commented.

Another unintended consequence has been an uptick in the length of index hospitalizations due to the presumption, which is not upheld by studies, that taking extra time to arrange for homecare, rehabilitation, and the like might mitigate the need for readmission, noted Sorin J. Brener, MD, of New York Methodist Hospital (Brooklyn, NY), in a telephone interview with TCTMD.

PCI a Minor Player?

Clearly, some readmissions are appropriate, whether planned or not, Dr. Spertus said. The CMS campaign has underscored the need not only to understand the reasons for readmissions but also to discriminate between those that can and cannot be prevented.

“For acute MI, I think in general the perception is that readmission rates are a poor measure of quality and almost entirely unrelated to the quality of coronary intervention,” James C. Blankenship, MD, of Geisinger Medical Center (Danville, PA) told TCTMD in a telephone interview.

The weak link between PCI quality and readmission rates has been underlined by findings from several studies.

For example, an analysis of 3,255 patients who underwent PCI at Dr. Blankenship’s institution found that 8.0% were admitted to the same or another hospital within 30 days. In only 11.9% of these cases—0.9% of all patients—was the readmission attributed to complications of PCI. Cardiac causes related to the index admission (but not to PCI) accounted for about one-third of readmissions, while the rest were due to noncardiac causes or factors completely unrelated to the original admission.

In another recent study, 9.8% of 9,081 PCI patients who were discharged alive were readmitted within 30 days. But only 6.7% of those readmissions were due to PCI complications, such as access-site bleeding or stent thrombosis, while 6.6% of patients returned for staged PCI without new symptoms.

For Interventionalists, It Comes Down to Good Care

In general, readmission rates after acute MI are not considered a core issue by interventional cardiologists, Dr. Blankenship said, because evidence shows that the quality of PCI—the factor most directly in their control—has relatively little bearing on readmission. Nonetheless, he qualified, interventionalists are committed to ensuring that their patients have a good transition from hospital to outpatient care and receive appropriate follow-up.

Dr. Spertus went further, saying: “The idea of interventionalists being responsible only for opening up the artery is faulty. They have to think of the patient’s whole care plan. With [the Medicare program], doctors are not really being penalized; it’s the hospitals that are being held accountable.” But as hospitals become more invested in reducing readmissions, “they will create greater alignment with physicians,” he predicted.

Similarly, Dr. Wasfy observed that “there are lots of ways that improving hospital systems can improve the ability of cardiologists to interact with their patients.” Examples are enabling timely communication via phone and/or video, he said, or alerting cardiologists when their patients show up in the ED.

An ‘Actionable Metric’

“I think hospital readmissions are an actionable metric, with significant opportunities for improving patient safety, reducing unnecessary cost, and encouraging providers to focus more on this measure,” Dr. Wasfy said. Yet the ability to identify effective strategies and target them to the patients who are likely to benefit most is both critical and challenging, he added.

A 2014 study he coauthored found that, at 2 Massachusetts hospitals, 42.6% of post-PCI readmissions were deemed preventable, with 1 out of 10 considered “definitely preventable” by 2 independently reviewing physicians.

The leading cause of preventable readmissions was staged PCI without new symptoms (14.7%), followed by vascular/bleeding complications, congestive heart failure, chest pain or angina-like symptoms, and stent thrombosis. After multivariable adjustment, the only factors associated with preventable readmission were diabetes, estimated glomerular filtration rate, prior PCI, and elective index PCI.

Although the largest proportion of preventable readmissions (34.2%) could have been avoided by performing an elective procedure at a different time, the researchers suggested that other readmissions could have been averted by:

• Different medical management (28.9%)
• Better access to outpatient care (10.3%)
• Performing a procedure differently (10.3%)
• Improved communication (8.4%)
• Prevention of vascular access complications (7.9%)

Among the readmissions due to stent thrombosis, MI, vascular/bleeding complications, or repeat revascularization, almost 10% were considered at least possibly preventable by procedural improvements. Still, the authors concluded that the majority could not be prevented with simple changes in decision making.

But Amy Boutwell, MD, MPP, founder of Collaborative Healthcare Strategies (Boston, MA), stressed that the definition of an unavoidable readmission “is largely in the eye of the beholder.” As clinical teams adjudicating readmissions gain experience and follow patients over time, they often develop a different perspective that leads them to view most readmissions as avoidable, she said in a telephone interview with TCTMD.

In an article about managing post-PCI readmissions, Dr. Blankenship and coauthors say both procedural and hospital complications can be sources of preventable readmissions. Procedural factors—such as vascular complications and coronary dissection—may be minimized by optimal management of individual patients, which may include choosing the safest access site and the appropriate stent type and size and tailoring antithrombotic therapy, they suggest. On the hospital side, early identification of patients at risk for nosocomial infections can alert the healthcare team to be more aggressive in following guideline-recommended preventive measures, they observe.

Predicting Likely Returnees

Of particular importance is identifying patients during the initial hospitalization who are more likely to be readmitted, Dr. Wasfy said. By matching the intensity of interventions to individual risk, he added, hospitals will reap a greater return on their considerable investment of money and organizational effort.

Toward that end, another study by Dr. Wasfy and colleagues used data from PCI patients treated in Massachusetts hospitals to develop and validate 2 multivariable models for predicting readmission within 30 days. The first incorporated only those variables known before PCI, while the second included variables known at discharge, with the latter only modestly improving upon the predictive ability of the pre-PCI model.

The HOSPITAL model, created by another set of researchers, recognizes 7 factors that predict readmission:

• Hemoglobin at discharge
• Oncology service discharge
• Sodium level at discharge
• Procedure during the index admission
• Index Type of admission
• Admissions during the last 12 months
• Length of stay

However, potentially important predictors such as functional status, health literacy, degree of social support, and previous medication compliance were not included, the authors acknowledge, because their goal was to produce a tool that could be used easily in the hospital.

Finally, in a registry study of more than 15,000 patients who underwent PCI at a single hospital, multivariate analysis found several patient traits and clinical and procedural factors associated with readmission, but most were not readily modifiable (eg, female sex, less than a high school education, and chronic diseases).

In an email with TCTMD, Robert W. Yeh, MD, MSc, of Massachusetts General Hospital (Boston, MA), said, “Predictive models for readmission all suffer from having a limited ability to discriminate [among] individual patients, but they do help to identify broad groups of high- or low-risk patients. They should include nonclinical factors, since those are some of the most important predictors of care access and subsequent readmission.”

The best application of the models is “selective targeting of high-risk groups for interventions that may be more expensive—things like early follow-up, visiting nursing care, or care navigators. These types of interventions are probably not cost-effective for all patients but may be very useful in these patients,” he commented.

In part 2 of this series, to be published on July 27, 2015, TCTMD looks at recommendations for reducing readmission and issues of how CMS should determine penalties.

Disclosures:

• Drs. Blankenship, Boutwell, Brener, Wasfy, and Yeh report no relevant conflicts of interest.
• Dr. Spertus reports being the founder of Health Outcomes Sciences.

Related Stories:

• Risk Model Predicts 30-Day Readmissions After PCI
• Hospital Readmission Post-PCI Linked to Increased Mortality
• Hospital Readmission After PCI ‘High’