Faulty gauge?

The Obama administration and health policy experts have been touting last year's decline in Medicare 30-day hospital readmissions as evidence that delivery and payment reforms designed to prevent unnecessary repeat visits were starting to succeed.

But one reason the readmission numbers may be declining is that hospitals increasingly are handling patients on so-called outpatient observation status, which in many cases is indistinguishable from inpatient admission. Since observational-status patients aren't counted as admissions, they aren't counted as readmissions if those patients are hospitalized within 30 days. Similarly, if observational patients had been hospitalized within 30 days prior to the observational-status treatment, that treatment wouldn't be counted as a readmission.

The possible interaction of the higher rate of observation status cases and lower rates of readmissions within 30 days is prompting tough questions about whether policymakers and providers should rely so heavily on 30-day readmissions as a key barometer of hospital quality and cost performance.

“Fundamentally, the question to me is, have we really done a good job of preventing readmissions, or have we just reassigned people who would have been readmitted to a different status?” said Dr. Ashish Jha, professor of public health policy at the Harvard School of Public Health in Boston.

The concern about the validity of Medicare's 30-day readmissions data comes just as the stakes for hospitals are rising. Last October, the CMS began docking hospitals' pay by up to 1% of total Medicare funding based on how many patients returned for care within 30 days for heart failure, heart attack and pneumonia, based on a program to cut preventable readmissions in the Patient Protection and Affordable Care Act.

Those penalties are set to double this fall, and then triple in October 2014. The CMS is also expected to add readmissions for more conditions in 2014, and is said to be considering extending the readmissions penalties to skilled-nursing facilities.

Outpatient observation is a form of non-acute care delivered to patients in hospital beds, and its use has exploded in the past five years—growing so fast that it may explain the decline in readmissions. In 2011, the number of Medicare outpatient observation cases in hospitals rose by 230,000 claims, according to an analysis of CMS data performed by the American Hospital Directory at the request of Modern Healthcare. Observation figures for 2012 are not yet available.

Experts say the move toward greater use of outpatient observation is due to the high rates of Medicare payment denials and aggressive auditing on short hospital stays.

Meanwhile, a study published last month in the Medicare and Medicaid Research Review—an online journal published by the CMS—found that the percentage of Medicare patients readmitted for any reason within 30 days declined to 18.4%, down from 19% in the previous five years. That translated into a decline of about 70,000 fewer cases in 2012, as would have been seen if the five-year trend had continued for a sixth year.

Preventable readmissions are a costly problem for the CMS. Nearly 1 in every 5 Medicare patients returns to the hospital within 30 days for additional care, costing the Medicare hospital trust fund about $18 billion a year. Some portion of those rehospitalizations is considered preventable though improved post-discharge care and coordination, though experts have strong disagreements about how much can be prevented, and by whom.

In a four-year study of hospital readmission-reduction programs in Massachusetts, Michigan and Washington state, researchers at the Institute for Healthcare Improvement in Cambridge, Mass., found that when total admissions drop, the number of readmissions declines as well. But the way the information was collected prevented researchers from quantifying directly how much impact observation rates had on readmission rates. “It's a contributing factor,” said Patricia Rutherford, a vice president with the institute and co-investigator on the study.

Here's how the observational-status and 30-day readmissions issues may interact.

Amy Deutschendorf, senior director of utilization and clinical resource management at Johns Hopkins Health System, Baltimore, offered the example of an 85-year-old patient with coronary stents and long cardiac history who comes to the hospital with fainting spells known as “syncope.”

Despite this patient's risk for heart arrhythmias and the signs of a symptom that he may be on the verge of serious cardiac event, hospitals are more likely to admit him to outpatient observation or treat him in the emergency department than admit him as an inpatient for syncope—even though the battery of testing he would undergo would likely be the same in any event.

“We can't admit that patient,” Deutschendorf said. “We would have three years ago.”

If the man were admitted to the hospital in the 30 days before or after the visit, putting him in observation could make the hospital look better on paper—even though the observational-status classification could have financial and quality-of-care repercussions for the patient.

No one necessarily accuses hospitals of consciously sending patients into observation as a way to avoid 30-day readmissions penalties. Whether the move was triggered by high rates of Medicare payment denials and aggressive auditing on short hospital stays or some other factor, hospital officials doubt it is a premeditated strategy.

Dr. Mark Williams, chief of hospital medicine for Northwestern University Feinberg School of Medicine, said he would be surprised if the trend was the result of any deliberate plan. “I'm sure there is someone in the U.S. who is doing it, but I've not seen cases of people who are talking about it,” he said.

Nancy Foster, vice president for quality and patient safety for the American Hospital Association, said there does appear to be some statistical relationship between falling readmission rates and rising use of observation. It would be unfortunate, she said, if Medicare readmissions penalties on hospitals were causing some clinicians to use observation status more. But she said that unintended consequences can happen even with well-intentioned metrics that are designed around broad goals that are supported by the evidence.

“When we get down to the nitty-gritty of a particular set of measures being applied in a particular way, that's when you discover that no measure is perfect,” Foster said. “And you may be invoking penalties and pressuring hospitals and clinicians to make changes that the science shows are not in the patients' best interest. And that is the tricky thing.”

Even if readmission rates may have fallen last year in part due to rising use of observational status, she added, that doesn't mean hospitals haven't made important strides in re-engineering their care processes to avoid preventable readmissions.

The CMS declined to provide comment for this article.

Meanwhile, 14 Medicare beneficiaries or their estates are suing the CMS in U.S. District Court in Connecticut, saying Medicare policies are encouraging hospitals to put patients in observational care inappropriately, and sometimes retroactively.

The decisions force patients to pay 20% Medicare Part B copays and the cost of prescriptions, plus the full cost of rehabilitation care, which ran upward of $30,000 for some of the plaintiffs. Patients who are admitted, and thus covered under Medicare Part A hospitalization services, have their rehab care paid for by the government.

Other elderly patients can't afford the high costs for rehab care that come with observation, and therefore they may forgo the necessary skilled nursing, said Ali Bers, an attorney for the litigants at the Center for Medicare Advocacy. “The lack of ability to get needed follow-up care can affect people's health,” she said.

A judge is expected to rule soon on whether to grant the CMS' motions to dismiss the lawsuit, and also whether to grant the case class-action status for seniors across the country affected by observation decisions.

Follow Joe Carlson on Twitter: @MHJCarlson