August 01, 2014
14 min read

Hospital readmissions present a complex, costly issue for US cardiologists

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Hospital readmissions for patients with heart disease are common and costly in the United States. Depending on the study population, research suggests that 20% to 40% of patients hospitalized for HF and 5% to 15% of patients hospitalized for ACS will be readmitted to the hospital within 30 days of discharge.

Harlan M. Krumholz, MD, is a leader in research focused on reducing the risk for hospital readmissions.

Harlan M. Krumholz, MD, is a leader in
research focused on reducing the risk
for hospital readmissions.

Source: Terry Dagradi;
photo printed with permission

Especially for patients with HF, the post-discharge period is one fraught with danger. “Patients hospitalized for HF are at intensely high risk,” Harlan M. Krumholz, MD, professor of medicine and director of the Center for Outcomes Research and Evaluation at Yale University, said in an interview. “There is no other group that has a risk of being hospitalized within any 30-day period of one in four.”

This has tremendous implications for patient care and health care costs. The US government has prioritized lowering readmission rates to the extent that CMS now issues financial penalties to hospitals with excessive 30-day readmission rates for patients with HF, acute MI or pneumonia. The Patient Protection and Affordable Care Act of 2010 also created new incentives to reduce hospital readmissions; hospitals with high rates can lose up to 3% of their Medicare reimbursement by 2015.

Unfortunately, lowering readmission rates is easier said than done. Efforts in the past decade to lower rates by improving patient adherence to post-discharge medication regimens and other instructions have mostly not paid off. Further complicating matters, in the majority of cases the patient is readmitted for a reason unrelated to the original admission. Another factor may be that efforts to reduce length of stay could have unintended consequences for readmission rates by encouraging discharge of some patients before they are ready.

Richard S. Schofield, MD, FACC

Richard S.

On top of that, many patients have a variety of comorbidities that make recovery more challenging and, especially in the case of HF, have disease states that can be very different from one another, making a single solution for every patient unlikely.

“I don’t think a magic bullet exists,” Richard S. Schofield, MD, FACC, professor of medicine in the division of cardiovascular medicine at the University of Florida, Gainesville, told Cardiology Today.

Causes, prevention of readmissions

The numbers paint a very complex and heterogeneous picture of readmissions in this patient population.

A study of patients with HF at the Mayo Clinic and other hospitals in Olmstead County, Minn., published in 2009 found that 16.5% of patients readmitted within 30 days were readmitted for HF and 21.6% were readmitted for other CV causes. In another study, aside from HF, the most common causes for readmission within 30 days for patients with HF included acute MI, cardiac dysrhythmias, chronic obstructive pulmonary disease, coronary atherosclerosis, fluid and electrolyte disorders, hypertension with complications, pneumonia, respiratory failure, and septicemia.

Matthew T. Roe, MD, MHS

Matthew T. Roe

Likewise, patients with MI who are readmitted within 30 days usually have to return to the hospital for a reason other than recurrent MI, said Matthew T. Roe, MD, MHS, associate professor of medicine with tenure in the division of cardiology at Duke University Medical Center and Duke Clinical Research Institute.

“It is very important in terms of readmission rates to focus on non-CV comorbidities because in the acute MI population many of the readmission reasons are related to non-CV reasons such as infections or other comorbidities that may exist in the patient,” Roe said in an interview.

Gregg C. Fonarow, MD

Gregg C.

The heterogeneity of causes for readmission means there is no consensus on how to prevent these rehospitalizations, as well as no consensus on to what extent rehospitalizations are preventable.

“While there are certainly some preventable rehospitalizations that occur, the extent to which they are preventable is very much in debate,” Gregg C. Fonarow, MD, co-chief of the division of cardiology at UCLA Medical Center, told Cardiology Today. “The percent to which they are related to the condition for which the patient is initially hospitalized is also very much in debate and greatly varies.”


Changes in thinking

It had been assumed by many cardiologists and hospital administrators that high readmission rates were a function of patient nonadherence and lack of patient education. But in reality, the problem is much more complex.

“We attributed this issue of readmission among patients with HF largely to patient noncompliance, dietary indiscretions and medication nonadherence to the point that we created an entire national quality structure of improving readmission rates in HF entirely based on that paradigm,” said Javed Butler, MD, MPH, professor of medicine at Emory University. “Now we have been at it for 10 to 15 years, and there is not a single measure associated with reducing the rehospitalization rate even 1% individually.”

Javed Butler, MD, MPH

Javed Butler

Image: Marty

According to Butler, a study by Bradley and colleagues of hospital strategies to reduce 30-day readmission rates for patients with HF found six strategies that improved rates, but with a modest magnitude of effect. Implementing them all could be very expensive and not make much impact, Butler said.

In the 1990s and early 2000s, CMS implemented a quality measure that specified instructions to give patients at discharge. “That translated into a check-the-box mentality, that you just print out something and give it to the patient,” Butler said in an interview. “But in spirit, nobody was teaching. You gave the patient a phone number to make an appointment, and that allows you to check a box. Now, we make sure to give the patient an appointment and follow up over the phone.”

The problem, Butler said, is that an approach focused on patient adherence and physician checklists ignore the biology of how HF occurs and worsens, and how HF actually consists of many different phenotypes. Likewise, those who think an existing drug regimen is the only answer are also missing the mark.

“In reality, to make a real dent in the readmission rate, we have to start [dividing] the population into different pathophysiologic segments and treating them according to which segment they belong to, rather than try to figure out one overarching intervention that would reduce the readmission rate,” Butler told Cardiology Today. “Until we get to the bottom of this, we will keep spinning our wheels putting in these generic interventions, which are very expensive. Hospitals employ many full-time equivalents that cost hundreds of thousands of dollars that can perhaps be better spent for other, more productive uses. We now have nearly a decade’s worth of experience that the current version of interventions, whether it is self-care education or telemanagement or discharge planning, are not improving outcomes to the extent either hoped or needed.”

If a patient with HF is readmitted, chances are one of three things happened, according to Butler. One possibility is that the patient stopped taking his medications or otherwise demonstrated extreme nonadherence. “It is important to remember that some degree of nonadherence with self-care recommendations are extremely common in stable outpatients with HF as well,” he said. A second is that the treating doctor did not aggressively treat the patient with guideline-recommended therapies or did not address comorbidities such as diabetes and arrhythmia disorders. The third possibility, however, is that the patient has worsening HF despite the doctor and patient doing what they are supposed to do. “This is where the biology is failing and the current medicines are not good enough,” he said. “This is especially the group for whom we need to find new targets and develop new drugs.”


The problem should also be looked at in the context of the entire health care system, Roe said.

“Looking at hospitalization as a single problem is a myopic viewpoint,” Roe said. “It is really a broader issue within the context of the entire health system and includes issues such as: Does the patient have insurance; does the patient have the means to pay for medication; is the patient able to take care of himself at home and follow the proper diet instructions; can the patient afford to participate in cardiac rehabilitation to maybe reduce the risk for readmission? All these things are intertwined.”

Indeed, in some contexts, rehospitalization may not be negative. A study of patients with HF in the Veterans Administration Health Care System between 2002 and 2006 found that 30-day readmission rates went up during that time, but 30-day mortality rates declined.

Length of stay as a factor

In recent years, research has explored whether there is an inverse relationship between 30-day readmission rates and length of stay and, if so, whether policies that encourage shorter lengths of stay are having unintended consequences of increasing readmission rates.

Krumholz and colleagues conducted an observational study of Medicare fee-for-service hospitalizations for HF from 1993 to 2006 (see Table). During that time, mean length of stay dropped from 8.81 days to 6.33 days, in-hospital mortality declined from 8.5% to 4.3% and 30-day mortality decreased from 12.8% to 10.7%, but 30-day readmission rates rose from 17.2% to 20.1% (P<.001 for all).

The relationship between the two metrics held across countries in the ASCEND-HF study, which found that countries with longer length of stay for HF hospitalizations had lower 30-day readmission rates. Moreover, in February, an analysis was published estimating that for every 1-day increase in length of stay, 30-day readmission rates were reduced 7% to 18% in patients with MI and 1% to 8% in patients with HF.

Table. Length of stay and short-term outcomes for Medicare patients hospitalized for HF

“The longer you keep people in the hospital, the less likely they’re going to come back within 30 days,” Schofield said. For patients with HF, “it is probably because we are getting their volume status more optimal the longer we keep them in the hospital. But there are tremendous economic pressures to get them out sooner. That’s the dilemma we face.”

Butler said new data currently being reviewed for publication from the EVEREST study will shed more light on whether there is a difference in the length of stay/readmission rate dynamic for patients with HF depending on cause of readmission: HF, other CV reasons or non-CV reasons.

However, a study by Robb D. Kociol, MD, and colleagues indicated that the 30-day readmission rate was poorly correlated with length of stay, quality measures and 30-day mortality for patients with HF, and that total inpatient days during a 30-day event of care may be a better metric to target for quality improvement efforts.

The study by Krumholz and colleagues “created this impression that it must have been the declining length of stay that has contributed to the higher rehospitalization rates,” said Fonarow, who was a researcher for the study led by Kociol. “But, in fact, if you look at the total days spent in the hospital during that time period — the days during the index hospitalization plus the days during readmission — the total days in hospital had actually fallen.

“Did the whole system benefit over the last 20-year period substantially by reducing the total economic burden due to HF? Yes. This impression that shortening the length of stay led to all these readmissions and is a major problem we should be focusing on is a little misleading,” Fonarow said.

The question that remains, however, is when is the best time to discharge a patient. This is especially difficult for patients with HF, in whom improved symptoms do not necessarily mean recovery, Schofield said.


“Clinicians are sometimes encouraged to let people leave before their volume status is really corrected,” he said. “A lot of times, patients will feel better with a partial correction of their fluid overload. The trouble is, if you let them leave when you’ve only partially corrected the problem, you’re setting them up for failure once they get home, because they’re just going to retain more fluid and come right back again.”

The problem, Schofield said, is that existing methods for determining whether fluid status is optimal are subjective and not easy to standardize.

Post-hospital syndrome

Given that many patients hospitalized for CV issues return to the hospital within 30 days, and that many return for reasons unrelated to their first visit, experts have questioned whether there is something about hospitalization alone that makes patients vulnerable to readmission.

In a January 2013 perspective in The New England Journal of Medicine, Krumholz hypothesized that newly discharged patients suffer from post-hospital syndrome, which he described as “an acquired, transient period of vulnerability.”

In post-hospital syndrome, “the risks in the critical 30-day period after discharge might derive as much from the allostatic and physiological stress that patients experience in the hospital as they do from the lingering effects of the original acute illness,” he wrote.

“I think that what happens is, people come in with a condition, they are put on bed rest, their circadian rhythms are thrown way off, they are sleep deprived and nutritionally deprived, they are given medications which can perturb their thinking or balance, and they are exposed to a lot of stress,” Krumholz told Cardiology Today. “They leave with almost a posttraumatic stress disorder.”

In this condition, patients may be cognitively impaired, may be at high risk for falls and other accidents, may have acquired an infection from the hospital, and may have impaired immune, renal or metabolic systems, Krumholz said. Therefore, it should be no surprise that they are susceptible to be readmitted for a different serious condition or to fail to remember the steps they are supposed to take to aid in their recovery, he noted.

“We should continue to talk to them about how to manage the condition they came in with, but especially if it was a chronic condition like HF, we can’t focus only on that, and we have to recognize the limitations that they’ve acquired during the course of the hospitalization,” he said in an interview.

In May, Krumholz and Allan S. Detsky, MD, PhD, from the University of Toronto, published a viewpoint in JAMA with suggestions for how to reduce the trauma of hospitalization. Strategies included ensuring that patients have adequate rest and nourishment; taking steps to reduce stress, disruptions and surprises; eliminating unnecessary tests and procedures; encouraging physical activity; and making follow-up appointments before discharge.

New solutions

Despite the complexity of this issue, a number of potential solutions are being tried at hospitals around the United States, although few methods have been proven by any research and even fewer by randomized clinical trials because of the financial penalties now imposed for high readmission rates.

One measure that has been associated with positive outcomes is follow-up within 7 days after discharge. Adrian F. Hernandez, MD, MHS, and colleagues conducted an observational study of patients aged 65 years and older with HF who were discharged between 2003 and 2006 from hospitals participating in the American Heart Association’s Get with the Guidelines–Heart Failure quality improvement program and in the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure. According to their results, patients discharged from hospitals with higher rates of follow-up within 7 days had a lower risk for 30-day readmission.

“That is an example of data generated by a quality improvement effort that led to new metrics that more and more hospitals are now striving for,” said Fonarow, who was a researcher of the study. “Medicare has even provided a framework for those early post-discharge visits to be in a special category.”


In many cases, that practice has been adopted for patients discharged with MI, according to Roe.

“Most integrated health systems, like our own, will ensure that MI patients have a visit within 7 days after discharge with a cardiology provider and will drill down carefully on their finances and their ability to procure and pay for their medication,” he said.

Marc Semigran, MD, medical director of HF and cardiac transplantation at the Massachusetts General Hospital Institute for Heart, Vascular and Stroke Care, said the hospitals-to-homes transition clinic at his institution includes similar practices.

“Patients are identified prior to discharge as having HF as a likely primary diagnosis and are set up to be seen in-clinic within 7 days of discharge,” he said. “Then they are followed in-clinic within 30 days of discharge, either with repeat visits, if they need them for patient education or titration of medications or other aspects of care, or via the phone, where they are followed by a nurse practitioner who knows them, has access to their medical record and can titrate their medications to optimize their health status and hopefully avoid readmissions. The initial results are that patients seen in the Hospital to Home transition clinic had a reduction in their 30-day readmission rate by one-third.”

Marc Semigran, MD

Marc Semigran

Also helpful is making sure patients are getting appropriate guideline-directed medical therapies, Semigran said.

“We have put into place ‘hard stops’ when clinicians write orders on patients with HF to make sure they are using guideline-directed medical therapies,” he said. “For example, the use of non-guideline–directed beta-blockers can be a problem. Similarly, isorbide dinitrate and hydralazine for African Americans with HF is grossly underutilized. We also have a natural language search program that we use to make sure we are identifying appropriate patients for resynchronization therapy and implantable cardioverter defibrillators.”

In fact, Roe said, patients at Duke discharged after acute MI, regardless of a stent, receive a free 30-day supply of antiplatelet therapy, either clopidogrel, prasugrel (Effient, Eli Lilly) or ticagrelor (Brilinta, AstraZeneca).

Crucial to the success of these programs will be rethinking of the stages of a patient’s illness, according to Krumholz.

“Strengthening what happens with systems when people leave is critical,” he told Cardiology Today. “One of the problems is that doctors are trained to think in segments. We were trained in the hospital to get people out of the hospital. We were trained in outpatient clinics to receive patients from the hospital. But for patients, these are episodes of illness. They don’t end when they go home. They continue over a period of time.

“We have to create systems where people leave the hospital into something that’s catching them. In that kind of world, I think you can shorten length of stay so they only spend time in the hospital when they really need acute care services, and then they leave into an environment where they’re not thrown out, but rather handed off into a system that acknowledges they are in the midst of the same episode of illness, they are at high risk, they could get worse again, and the goal is to protect them from that,” Krumholz said.

Addressing the issues in context

Although it appears that practices that could improve 30-day readmission rates are being implemented all over the United States, experts caution that it is too early to tell whether they can be effective and, if they are effective, it is too early to tell whether they will lead to improved long-term outcomes for patients.

“Until further data come out, we should not equate [success in reducing readmission rates] with improved outcomes,” Butler said. “Until we find out what happened to the patient and the HF, we cannot declare victory. Even if due to financial concerns we lower the readmission rate by shifting care to alternate venues such as observation units, nursing homes or outpatient infusion centers, this reduction in readmission rate should not be considered as improving the outcomes of patients with HF. In these cases, the patient hasn’t necessarily won. We have to be careful how we define that.”

Therefore, some say, readmission strategies should be incorporated into an overall strategy of care, and not be considered in a vacuum.

“An overly incentivized approach to just focusing on 30-day readmission is too narrow and short-sighted,” Fonarow said. “A more comprehensive approach to better management of patients in all phases of care and with a key goal of preventing the index hospitalizations is really critical. We need to better manage outpatients with HF to prevent them from needing hospitalization in the first place and to manage patients with or at risk for coronary disease so they are not being admitted with acute MI.” – by Erik Swain

Blecker S. J Am Coll Cardiol. 2013;61:1259-1267.
Bradley EH. Circ Cardiovasc Qual Outcomes. 2013;6:444-450.
Bueno H. JAMA. 2010;303:2141-2147.
Carey K. Med Care Res Rev. 2014;71:99-111.
Detsky AS. JAMA. 2014;311:2169-2170.
Dunlay S. J Am Coll Cardiol. 2009;54:1695-1702.
Eapen ZJ. Circ Heart Fail. 2013;6:727-732.
Heidenreich PA. J Am Coll Cardiol. 2010;56:362-368.
Hernandez AF. JAMA. 2010;303:1716-1722.
Kociol RD. Am Heart J. 2013;165:987-994.e1.
Krumholz HM. N Engl J Med. 2013;368:100-102.

Javed Butler, MD, MPH, can be reached at Emory University Hospital Midtown, 550 Peachtree St., Atlanta, GA 30308; email:
Gregg C. Fonarow, MD, can be reached at 10833 Le Conte Ave., Los Angeles, CA 90095; email:
Harlan M. Krumholz, MD, can be reached at 1 Church St., Suite 200, New Haven, CT 06510; email:
Matthew T. Roe, MD, MHS, can be reached at 2400 Pratt St., Durham, NC 27705; email:
Richard S. Schofield, MD, FACC, can be reached at 1600 SW Archer Road, Box 100277, Gainesville, FL 32610; email:
Marc Semigran, MD, can be reached at Bigelow 800, Massachusetts General Hospital, Boston, MA 02144; email:

Disclosure: Butler reports receiving research support from the European Union, the Health Resources Services Administration and the NIH and consulting for Amgen, Bayer, BG Medicine, Cardiocell, Celladon, Gambro, GE Healthcare, Medtronic, Novartis, Ono Pharma, Otsuka, Takeda, Trevena and Zensun. Fonarow does volunteer work for the American Heart Association’s Get with the Guidelines program and reports consulting for Amgen, Bayer, Gambro, Medtronic and Novartis. Krumholz is director of the Center for Outcomes Research and Evaluation at the Yale University School of Medicine, which has a contract from CMS to develop readmission-related measures. Roe reports financial ties with Amgen, AstraZeneca, Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly & Co., Elsevier Publishers, the Familial Hypercholesterolemia Foundation, Janssen Pharmaceuticals, Merck and Sanofi-Aventis. Schofield reports serving on an advisory board for Cytori Therapeutics. Semigran reports no relevant financial disclosures.