Medicare’s Bundled Payments Plan Hasn’t Cut Costs (Yet)

The 2013 BPCI initiative, with acute MI and CHF under its umbrella, hasn’t had an impact. That doesn’t mean healthcare delivery doesn’t need a rethink.

Medicare’s Bundled Payments Plan Hasn’t Cut Costs (Yet)

Attempts to save healthcare dollars in the United States by bundling Medicare payments for particular medical conditions such as acute MI on a per-episode basis haven’t yet worked, according to a new analysis of data from 2013 through 2015.

Medicare payments did not decrease among participating hospitals, nor did various measures of resource use. Clinical complexity and mortality rates also did not change, researchers reported today in the New England Journal of Medicine.

Launched in 2013, the Bundled Payments for Care Improvement (BPCI) initiative is voluntary. Hospitals that choose to participate can pick among 48 conditions and agree to be held accountable for care in time frames of 30, 60, or 90 days. If cost targets are met, the hospitals get to keep some of the savings. If not, they must reimburse Medicare for some of the extra expense.

Earlier data on BPCI and lower-extremity joint replacement showed a decrease in Medicare payments, suggesting that the same might be true for other medical conditions, lead author Karen Joynt Maddox, MD (Washington University School of Medicine, St. Louis, MO), told TCTMD. But joint replacement is an elective procedure that tends to be done in younger, less sick patients than those being treated for acute MI, congestive heart failure, and pneumonia. “We’re talking about people with multiple comorbidities and really quite a lot of need, a lot of nursing home use, presumably a lot of really complex disease,” she explained.

The results don’t show that bundled payments are a failure, Joynt Maddox said, pointing out that the median follow-up in the new study is under 1 year. Rather, they show that in terms of reducing costs, “there was not a quick and easy fix. That’s not to say that there’s not a slow and difficult fix, though,” she noted.

Building partnerships and achieving clinical change to “move the needle” will take time, Joynt Maddox said. There’s “incredible inertia” in healthcare systems, she observed. “We still deliver care largely the way we did 50 years ago. People come to their outpatient appointments and they come to the hospital, and we really haven’t reimagined care in a long time. So thinking that the bundled payment system would be the magic bullet—that there would be something easy that just no one was doing yet—[is unrealistic].”

Many physicians see the value in rethinking healthcare, Joynt Maddox pointed out. “As clinicians, we see how much better the system could be for our patients. We know we work in a bad system. We know it’s not patient-friendly. We know that the coordination isn’t good. We know we could do better. And the problem is until the payment structure changes, it’s really hard to get the collective will you need from a health system, an organization, [in order] to change. . . . In the absence of payment incentives that can get the CEOs and CFOs talking care redesign, a frontline clinician [alone] can’t change the system.”

It’s not that it doesn’t work, it’s that there’s not an easy fix. Because if there was one, people would have done it. This is going to be hard. Karen Joynt Maddox

Joynt Maddox and colleagues focused on the five most common medical conditions chosen by hospitals taking part in BPCI: congestive heart failure, pneumonia, chronic obstructive pulmonary disease, sepsis, and acute MI. The number of participating hospitals varied among the conditions, from 125 for congestive heart failure to 73 for acute MI, and they tended to be nonprofit, urban, teaching hospitals with a large number of beds. The researchers looked at the 3 to 9 months before each hospital started BPCI as well as the 3 to 9 months afterward, and they matched these periods to times at nonparticipating hospitals serving as controls.

The average Medicare payment per 90-day episode of care did not change significantly at either participating hospitals, where costs dropped by $286 (P = 0.41), or at controls, where the decrease was $398 (P = 0.08). Length of stay, emergency department use, and hospital readmission also were unaffected by BPCI participation. Additionally, patients had similar demographic characteristics and clinical complexity.

Thirty-day and 90-day mortality rates rose to a similar degree at both BPCI and control hospitals.

“Bundling of services to encourage more efficient care has great face validity and enjoys bipartisan support,” the researchers conclude. “For such bundling to work for medical conditions, however, more time, new care strategies and partnerships, or additional incentives may be required.”

Certainly, there has been ongoing interest in bundled payments. In the summer of 2016, the Centers for Medicare & Medicaid Services proposed an involuntary program that would apply lump-sum payments for MI care and bypass surgery, with incentives for cardiac rehab. That plan was finalized in late 2016, with the start date eventually set at January 2018. Before it could even begin, however, CMS reversed its course and cancelled the reforms.

Earlier this year, CMS announced the latest version of the BPCI plan, also voluntary, called BPCI Advanced. The strengths of BPCI Advanced are that “it’s simpler, it’s a little more streamlined, the feedback’s going to be better, and the risk adjustment’s going to be better,” Joynt Maddox said.

“What our study would suggest is that even as that program starts, we need to give it time,” she cautioned. “It’s not that it doesn’t work, it’s that there’s not an easy fix. Because if there was one, people would have done it. This is going to be hard.”

Caitlin E. Cox is News Editor of TCTMD and Associate Director, Editorial Content at the Cardiovascular Research Foundation. She produces the…

Read Full Bio
  • Epstein reports serving as an associate editor for the New England Journal of Medicine.
  • Joynt Maddox reports receiving personal fees from the US Department of Health & Human Services unrelated to this study.
  • Orav and Zheng report no relevant conflicts of interest.