No Hospital Cost Savings From Bundled MI, HF Medicare Payments

Debra L. Beck

July 25, 2018

In an analysis of data from 2013 to 2015, hospitals testing bundled payments for episodic care for five common medical conditions, including congestive heart failure (CHF) and acute myocardial infarction (MI), failed to lower their costs substantially.

Participating hospitals in the Center for Medicare and Medicaid Innovation (CMMI) Bundled Payments for Care Improvement (BPCI) received similar Medicare payments and showed about the same readmission and mortality rates compared with those not participating in the BPCI, researchers reported in their study published July 18 in the New England Journal of Medicine.

"This answers the question of whether or not there is a quick and easy fix to lowering costs, and the answer is no," said Karen Joynt Maddox, MD, MPH, Washington University School of Medicine, St. Louis, Missouri, in an interview with | Medscape Cardiology.

"But we don't want this to be seen as an indictment of the concept, but rather as a cautionary tale that just simply flipping a payment switch does not immediately produce results," she said, adding that she "would love to see this work" for many reasons, including that it will likely lead to more patient-centered care.

BPCI is a voluntary program designed to lower costs by enhancing transitions of care between acute- and post-acute care. Launched in 2013, the program hopes to develop cost-containing, episode-based payments for common medical conditions requiring hospitalization.

Claims for the Top Five

Joynt Maddox and colleagues focused only on claims related to the five most commonly selected condition in BPCI: CHF, pneumonia, chronic obstructive pulmonary disease (COPD), sepsis, and acute MI.

Researchers used 6 months before each hospital's start date as the baseline period and looked at billing over the next 3 to 9 months. These periods were matched to billings from nonparticipating hospitals that served as controls.

From a pre-intervention baseline average (for all five conditions) of $24,280, payments decreased to $23,993 during the intervention period for BPCI participants (P = .41).

Nonparticipating hospitals had baseline average payments for all episodes of $23,901, which decreased marginally to $23,503 during the intervention period (P = .08) and were not significantly lower (P = .79) than those seen in BPCI participants.

No significant differences were noted for either intervention or control hospitals from baseline through the intervention period in clinical complexity, length of stay, emergency department use, 30- or 90-day readmission, or 30- or 90-day mortality.

Under BPCI, hospitals are paid fee-for-service rates, but Medicare-allowed payments are retrospectively reconciled against target sums on a quarterly basis, the authors explain.

There are four different BPCI models, they continue, but most of the participating hospitals have chosen BPCI model 2, wherein the hospital assumes accountability for the costs of all acute- and sub-acute care within 30, 60, or 90 days after hospitalization for 1 of 48 conditions.  

A total of 125 hospitals participated in the program for CHF, 105 hospitals for pneumonia, 101 hospitals for COPD, 88 hospitals for sepsis, and 73 hospitals for acute MI. The volunteer hospitals chose the conditions and time window.

"This is going to take real work, but these are the common conditions that people are admitted to the hospital for and the drivers of healthcare costs that we need to be working on," Joynt Maddox said.

The analysis comes on the heels of a report published last month in JAMA Cardiology showing the hospitals that volunteered for the BPCI payments program were not representative of the nation's hospitals, but rather were larger, higher-volume facilities less likely to care for large numbers of Medicaid patients.

Joynt Maddox noted that the selected group of volunteer hospitals participating in BPCI are those most likely to succeed because of their size and strength, but the specifics of which hospitals received which payments and met goals have not been released to the public by the Centers for Medicare and Medicaid Services (CMS).

"What's hidden in this average of nothing is a pretty broad range of savings versus costs going up," she said. "We're reporting one number, but buried within that number are some success stories, so the other thing we need to do is start learning from those institutions to learn what works."

Policy Pros Rooting for Bundled Payments

On the study's publication, the health policy Twittersphere lit up with comments expressing appreciation for the quality of the research and disappointment with the findings.


Ashish K. Jha, MD, MPH, is an internist at the VA Boston Healthcare System and director of the Harvard Global Health Institute, Boston, Massachusetts. He further speculated that, based on this and other recent research, the evidence shows that bundled payments may be cost saving for surgical conditions but not medical conditions, and may not "do much to improve quality."

Other policy wonks responded with similar sentiments:




Joynt Maddox agreed that it's too early to draw conclusions on whether the concept might work, given enough time and adjustment.

"Our follow-up for many hospitals was only 3 or 6 months and it may take time to see the kind of transformation we need to really move the needle," she said.

"The evidence we have so far that bundled payments work mostly comes from joint replacement, which is a very different population than the folks coming into the hospital for heart failure, COPD, etc," she said.

"It may be that for these complex conditions and complex patients, we need more complex interventions and it's just going to take longer."


"Ideally, you would really want more people under the same financial risk umbrella — the nursing homes and outpatient providers, and everyone, tied together — if you really think that financial risk is what's going to move the needle, but our system is just so fragmented right now, so using the hospitals essentially as leverage to try to create some bridges across those settings I think is a reasonable approach," said Joynt Maddox.

She added that "with the administration doubling down on this" with BPCI Advanced, which will be launched at the beginning of fiscal year 2019, it's important to better understand how these new payment models are working and for whom.

The study was supported by the Commonwealth Fund. Joynt Maddox has reported no relevant financial relationships. Disclosures for the other authors are listed on the journal website.

N Engl J Med. 2018;379:260-269.

For more from | Medscape Cardiology, follow us on Twitter and  Facebook.



Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: