Getting a Gauge on Readmissions Penalties

In October, Medicare will begin tying payments to some of its performance measures. Hospitals that fare poorly on preventable readmissions for acute MI, heart failure and pneumonia may see reimbursement reduced by up to 1 percent in 2013. The penalty applies to all Medicare discharge payments during a fiscal year, and not just payments for preventable readmissions, making the financial loss potentially huge. Are you ready?

Nearly one in four Medicare beneficiaries admitted to a hospital with a diagnosis of heart failure (HF) between 2005 and 2008 was readmitted within 30 days (Circ Cardiovasc Qual Outcomes 2009;2[5]:407-413). There was wide variation across hospitals and regions, indicating a potential for improvement in performance and patient outcomes.

Readmissions also bear high costs. In 2008, the Medicare Payment Advisory Commission reported that Medicare spent $12 billion in 2005 on potentially preventable hospital readmissions, with HF readmissions contributing $903 million to the total. In an effort to curb these costs, the Patient Protection and Affordable Care Act (PPACA) established a hospital readmission reduction program that gave the Centers for Medicare & Medicaid Services (CMS) authority to penalize hospitals with greater than expected risk-adjusted readmission rates for HF, acute MI and pneumonia. Tracking is based on all-cause readmissions.

Beginning in fiscal year 2013, CMS may withhold up to 1 percent in inpatient Medicare payments to poor performers, increasing to 2 percent in 2014 and 3 percent in 2015. Also in 2015, CMS may expand the program to include chronic pulmonary obstructive disease, CABG, PCI and other vascular conditions. “For hospitals with high Medicare service volume, and a high preventable readmissions rate, the aggregate effect of the penalty could be considerable,” the Congressional Research Service wrote in the 2010 report, “Medicare Hospital Readmissions: Issues, Policy Options and the PPACA.”

Measuring up
The American College of Cardiology (ACC) and the American Heart Association (AHA) published performance measures for HF in 2005 to provide physicians with evidence-based guidelines defining practices that facilitate optimal patient care. With acute MI, there is strong evidence that high performance based on guideline recommendations improves outcomes, says Robert O. Bonow, MD, co-chair of the writing committee for the HF performance measures, which were updated in 2011. But with HF, whether hospitals comply with the recommended measures, or if compliance leads to improved outcomes, is less clear.

“In acute MI, [with compliance] you have lower mortality and recurrent events,” says Bonow, director of the Center for Cardiovascular Innovation at Northwestern University Feinberg School of Medicine in Chicago. “With heart failure, it is much more tricky because the connection between high performance among those processes of care—giving the right drugs, discharge instructions and so forth—and tying those to either a reduction in mortality or, most importantly, 30-day readmissions, that connection has been lacking.”

David W. Schopfer, MD, a cardiologist at the San Francisco Veterans Affairs Medical Center, and colleagues tried to clarify the connection by evaluating compliance among 3,655 hospitals identified through CMS’ Hospital Compare open-access database that admitted patients with HF diagnoses in 2008 (Am Heart J 2012;164[1]:80-86). For their analysis, they focused on four of the 2005 measures: evaluation of left ventricular systolic function; administration of angiotensin-converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARBs); providing smoking cessation counseling; and giving discharge instructions.

They found that socioeconomic status and hospital volume were stronger predictors of 30-day mortality and readmission rates than compliance, while only evaluation of left ventricular systolic function and smoking cessation counseling were associated with lower HF readmission rates.

“Hospitals need to focus on strategies beyond compliance,” says Schopfer, citing as an example his medical center’s protocol to contact the patient within 48 hours after discharge to ask about his or her status, review medications and answer any questions. “It is an opportunity to clarify in case there is a problem.”

But many hospitals may not implement such strategies. In a study designed to define the range and prevalence of recommended practices for reducing 30-day readmissions of patients with acute MI or HF, researchers found that on average hospitals had only 4.8 of the 10 in place (J Am Coll Cardiol online July 18). Only 12 percent had implemented eight or more practices, while another 12 percent implemented two or less.

“We are looking at what hospitals are doing to reduce readmission rates and finding tremendous variations in practices,” says lead author Elizabeth H. Bradley, PhD, director of the Global Health Leadership Institute and a professor at Yale School of Public Health in New Haven, Conn.

Beyond process measures
Bonow categorizes measures, such as administering ACEI or ARBs, as relatively simple processes of care that can be captured in administrative data. Other processes added to the 2011 guidelines, patient self-care education and arranging and documenting follow-up with a discharged patient, also provide opportunities to improve the quality of care. The 2011 updated guidelines reflect that shift with an emphasis on continuity of care between the inpatient and outpatient settings.

“These are the soft points one needs to start working on,” Bonow says.

But do these and other practices lower HF readmissions rates, and if so, which have the biggest impact? Bradley and her colleagues aim to answer those questions using their study as a baseline. In a continuation of the research, they plan to statistically correlate practices with risk-standardized readmission rates to identify the core practices associated with improved performance. Using a similar approach with acute MI, they found five strategies that were associated with lowered mortality rates: having physician and nurse champions rather than nurse champions alone; fostering a problem-solving environment; holding monthly meetings with hospital and patient transport staff; having a cardiologist on site 24/7; and not cross-training intensive care nurses for the cath lab (Ann Intern Med 2012; 156:618-626).

The mean 30-day risk-standardized mortality rate for hospitals that implemented all five practices was a full percentage point lower than for those that implemented only two. “We find the better hospitals that have a culture that is conducive to coordination and have [communication] systems built to facilitate the coordination are getting the best outcomes,” Bradley says. “It is not those hospitals that are investing the most in the bricks and mortar and equipment.”

Of those five strategies, Bradley places a problem-solving culture as the most impactful but also the most challenging. “When you problem solve, you first have to realize there is a problem and that is hard; many hospitals are in denial,” she says. Data analysis takes time, and there is a risk that the results may lead to a blame game rather than a constructive response. “It is culturally the most complicated, but it also is the most important.”

Challenges with HF
The care process for acute MI and HF differ, Bonow points out. Acute MI patients are more easily identified through processes that quickly steer them to the cardiac cath lab or a cardiac care unit. HF patients, on the other hand, may be admitted by an internist or other non-cardiac specialist rather than a cardiologist; they may be scattered throughout the hospital; and their condition may not be recognized as HF until several days after admission.

“Often a hospital knows about its heart failure patients only when it is retrospectively sending in administrative data for billing purposes,” he says. “Heart failure may not be the primary diagnosis for the admission even though it may be driving a lot of what is going on in the illness and patient care needs. The admission diagnosis may not evolve with the discharge diagnosis.”

The confounding relationship between readmissions and factors such as mortality and patient comorbidities adds to the challenge. For instance, in a 2010 analysis of Hospital Compare data, 19 of 35 hospital referral regions with worse than expected risk-standardized readmission rates for HF also ranked better than expected for mortality rates (Circ Cardiovasc Qual Outcomes 2012;3:459-476). Gorodeski et al reported similar trends in their analysis of Hospital Compare, and suggested that hospitals with lower mortality rates had a greater proportion of discharged patients who therefore may be readmitted (N Engl J Med 2010;363[3]:297-298). A more recent editorial argued that higher readmission rates may reflect hospitals’ capability to keep sicker patients alive or improved access to needed care (N Engl J Med 2012;366[15]:1366-1369).    

Also, the initial HF or acute MI may not be the reason for a readmission, Veronique L. Roger, MD, MPH, of the health sciences research department at Mayo Clinic in Rochester, Minn., and colleagues have shown. In a study that looked at the lifetime burden of hospitalization of patients after a first-time HF diagnosis, they reported that 83.1 percent of HF patients were hospitalized at least once over the 4.7-year mean follow-up. HF was listed as the reason in only 16.5 percent of the hospitalizations, while more than half were noncardiovascular (J Am Coll Cardiol 2009;54[18]:1695-1702).   

Calculating comorbidities
In a drill down of acute MI data in a population-based registry in one county in Minnesota, they found that one in five patients who experienced their first MI were readmitted to the hospital within 30 days, but only 42.6 percent of rehospitalizations were related to the index MI or its treatment (Ann Intern Med 2012;157:11-18). Inpatient survival improved between 1987 and 2010 and the frequency of hypertension, hyperlipidemia, diabetes, obesity, chronic obstructive pulmonary disease and anemia also increased over time.   

“Because those patients have this context of extensive comorbidities that is increasing over time, they do get readmitted to the hospital but for other reasons,” Roger says. Nonetheless, because MI was the initial diagnosis, the readmission would fall under the reduction program with potential financial consequences for the hospital.

Like acute MI patients, HF patients have a high incidence of multiple morbidities that increases over time. The proportion of HF patients with five or more comorbid conditions increased from 42.1 percent in 1988 through 1994 to 58 percent in 2003 through 2008 (Am J Med 2011;124[2]:136-143). Roger, Bradley, Bonow and Schopfer agree that to be effective, physicians need to evaluate the many chronic conditions and circumstances that affect patient care.

“These data underscore the need of focusing on the patient as a whole,” Roger says. “We need to have a holistic approach to care that is not only disease-focused but patient-centric.”
Bradley sees concepts involving communication and culture as a cornerstone for improving outcomes that may apply to HF or acute MI patients with or without multiple chronic conditions. Her team is developing actionable recommendations, such as sending a discharge summary electronically to the patient’s outpatient physician within 48 hours of discharge, as well as strategies that help to reinforce communication and coordination of care within the organizational environment.

They are collaborating with the ACC on a campaign to make these resources available within 2012. “Some of this involves pulling out very concrete behaviors,” Bradley says, such as giving caregivers discharge instructions or having a reliable system for identifying HF patients at the time of admission. “Others include more intergroup work and more clarity in roles, approaches that are related to human resources and how people manage together inside a group.”  

A focus on healthcare literacy, including efforts to ensure that a patient and his or her family understand what and how to maximize care once the patient is discharged, offer an opportunity to improve outcomes, according to Schopfer. This strategy may be useful for hospitals that serve urban and disadvantaged patient populations.

Some hospitals, such as Northwestern, have been developing care teams to focus on HF patients during their admission and discharge, as well as HF clinics to assist patients in the outpatient setting with post-discharge communications and support. “You can have an impact on 30-day readmissions, but because this is a population that is aging with comorbidities, there always will be some level of readmissions,” Bonow cautions. “What you are trying to do is identify and prevent the unnecessary readmissions, which are related to providing the best kind of continuity of care.”

Roger and Bonow see the emphasis on readmissions as an opportunity to expand beyond siloed care and instill a hospital-wide culture of safety and quality. Readmissions offer physicians a second chance to identify gaps in treatment that initially had been overlooked and fine tune the patient’s care.

Linking payment penalties to all-cause readmissions potentially may ding hospitals for non-HF related rehospitalizations, but it prevents hospitals from gaming the system and may give them an incentive to address the many needs of these complex patients at the initial admission.

“Some of that [penalty] is reasonable, because we should have taken care of the other comorbid conditions in the first place,” Bonow says. Better care may require a longer hospital stay, more dedicated staff or more educational resources but the offshoot may be an improvement in quality and safety. “Right now that may be costly for a hospital, but it may be well worth it if you can prevent the readmission.”