The data are in, but debate rages: Are hospital readmission penalties a good idea?

A closely watched experiment in health care has unfolded over the last few years: Financial incentives for hospitals to reduce readmissions. It was a feature of the Affordable Care Act, and researchers are now assessing its impacts. But even as data have arrived, an impassioned argument has broken out among experts — looking at the same numbers, they have reached different conclusions about whether the policy is making Americans healthier.

And, underscoring that disagreement is a deeper one — about what kind of evidence is needed before a health policy is enacted on a national scale.

The policy, known as the Hospital Readmissions Reduction Program, created financial penalties for hospitals whose readmissions exceed the national average for patients suffering from heart failure, heart attacks, and pneumonia. In recent years it has been expanded to include other conditions. Its aim was to encourage hospitals to deliver stepped-up care to severely ill patients even after they leave the hospital, in the hope of preventing return visits that result in more anguish for patients and skyrocketing costs for everyone else.

And one thing scholars generally agree on is that the program has achieved its primary goal: It has reduced readmissions. A wide body of evidence shows that readmissions began to fall in 2012, when financial penalties took full effect. They have since declined several percentage points in each of the three conditions originally included in the program, according to a Kaiser Family Foundation analysis of Medicare data.

But skeptics say those results belie a darker truth — that hospitals are taking shortcuts, and in some cases compromising patient care, to avoid financial pain and public embarrassment.

One recent report by the University of Michigan found that a large percentage of the reduction in readmissions is attributable to changes in the way hospitals are describing their patients in claims data. By describing them as sicker, hospitals can increase their risk adjustments, thus reducing financial penalties.

But the biggest flashpoint emerged last month when researchers at UCLA and Harvard published a study that correlated the reduction in readmissions with an increase in 30-day and one-year mortality rates among heart failure patients.

The study suggested that the program was backfiring for those patients, keeping them out of the hospital but thereby jeopardizing their health. And the result was more deaths. “If further confirmed,” the report concluded, “these findings may require reconsideration of the [readmissions program]” in heart failure.

Like most debates in science and politics, the one over readmissions is far from black and white. The research on its effects is colored by subjectivity and layers of nuance that make it difficult to establish the bottom-line truth of how it is impacting outcomes for patients.

And in this case, experts who have invested years generating and examining these data are in sharp disagreement over what conclusions can be drawn from the numbers.

Data on deaths sparks intense debate

Standing at the center of the argument is Dr. Harlan Krumholz, a cardiologist at Yale who directs its influential Center for Outcomes Research and Evaluation. The center’s research, conducted over many years, formed the basis for the readmissions policy that was eventually included in the Affordable Care Act.

Krumholz and his colleagues proposed the readmissions measure over a decade ago and devised how hospitals’ rates would be calculated and risk-adjusted to reflect their patient populations. The idea was to counteract a perverse incentive hospitals previously had: By providing poor follow-up care, they got increased revenue from more return visits to the hospital.

“The hospitals would tell me, ‘I can reduce readmissions, but that costs us money,’” Krumholz said. “There was a great inertia, and nobody was investing any money” to fix it.

Krumholz’s own research has found that the the readmissions policy is working. Earlier this year, he and his colleagues released a study that found the policy has resulted in reduced mortality at hospitals that have cut their readmissions.

However, the study, published in the Journal of the American Medical Association, also found that, when looking at aggregate data — rather than numbers at specific hospitals — the 30-day mortality rate among heart failure patients had ticked up by 1.3 percent nationally.

That latter finding sparked a public debate between Krumholz and Dr. Gregg Fonarow, a cardiologist at UCLA who was also examining the impacts of the readmissions program.

In response to Krumholz’s study, Fonarow and a colleague wrote a letter to JAMA to argue that the 1.3 percent increase was far from a side note — and instead represented “the worst case scenario” effect of the readmission program.

Krumholz disagreed, writing a response letter that warned of the “potential pitfalls” of drawing conclusions about the impact of the readmissions program by looking at aggregate data, which can be affected by contextual factors, such as broader changes in the patient population and the medical problems they face.

But in November, Fonarow and colleagues produced their own study with a more alarming conclusion: The policy was associated with deaths of patients. They found that the 30-day mortality rate — which calculates deaths within a month of hospitalization — increased from 7.2 percent before the program’s implementation to 8.6 percent after federal officials began assessing financial penalties. The study’s sample included more than 115,000 Medicare beneficiaries that participated in a registry of heart failure patients maintained by the American Heart Association. If extrapolated nationally, its findings mean thousands of additional deaths could be occurring each year as a result of the policy.

Fonarow said he and his co-authors could not pinpoint the reasons for the increase in mortality. But he suggested hospitals’ concerns about financial penalties could be causing them to lean against readmitting patients.

“You can imagine scenarios where patients were left at home or discharged straight from the emergency room when their … outcome would be better with hospitalization in a more closely monitored environment,” he said.

While Krumholz and Fonarow agree about the rates, they disagree about whether the readmissions policy is to blame. In the days after Fonarow’s study was released, they sparred over its conclusions on Twitter.

“No level of reduction of readmissions or cost savings should be considered adequate justification for a mortality increase,” Fonarow tweeted.

Krumholz replied, “Policies should be continually evaluated for unintended harm, but needs to be a modicum of evidence of harm,” he wrote in response. “If you think there’s harm, why aren’t (the American College of Cardiology and American Heart Association) immediately calling for doctors to stop harming patients by not admitting them [?]”

Bigger questions emerge

In recent weeks, doctors nationwide have jumped into the fray, tweeting their own thoughts and reactions — and asking whether the questions being raised about the readmissions program should have been answered before it was implemented on a national basis.

“I find there to be some cognitive dissonance in the way those who call for the absolute highest levels of evidence for treatments such as device and drugs (which I agree with) are often willing to relax those standards in the evaluation of health policy,” said Dr. Robert Yeh, a cardiologist at Beth Israel Deaconess Medical Center, in an email. “Bad health policies, like bad drugs, have unanticipated side effects.”

Last week, Yeh and his colleagues published another study focused on readmissions. It zeroed in on a different subset of patients — those with peripheral arterial disease who undergo procedures to restore blood flow through blocked arteries.

The study found that such patients are readmitted at a rate of 17.6 percent. Each such hospitalization costs an average of $11,000 and leaves patients vulnerable to further complications — impacts that could warrant the inclusion of these patients in the national readmissions program.

However, Yeh and his co-author, Dr. Eric Secemsky, also found that differences in hospital quality only marginally accounted for the patients’ differences in readmission risk, which raises the question of whether financially penalizing hospitals makes any sense.

Yeh said the only certainty at this point is that the readmissions program has focused more attention on the need to provide better follow-up care. “The jury is still out on whether the (program) has somehow created harm,” he said. “We can say with much more confidence that it has kept many patients out of the hospital.”