Participation in Medicare’s bundled payments initiative didn’t significantly change payments per episode or care outcomes for the top five medical conditions selected under the program, a new analysis shows.
Payments for the common conditions remained around $24,000 per episode before and during participation in the Bundled Payments for Care Improvement (BPCI) initiative for the 125 participating hospitals evaluated in this study, conducted by Karen E. Joynt Maddox, MD, of Washington University, St. Louis, and her coauthors.
The finding contrasts with a previous study showing that hospitals in BPCI successfully lowered overall Medicare payments for patients who underwent joint replacement.
“Bundling of services to encourage more efficient care has great face validity and enjoys bipartisan support,” Dr. Joynt Maddox and her colleagues wrote. “For such bundling to work for medical conditions, however, more time, new care strategies and partnerships, or additional incentives may be required.”
The Center for Medicare & Medicaid Innovation initiated the voluntary BPCI demonstration project in 2013. The program targets 48 conditions that account for about 70% of Medicare spending. Hospitals that achieve cost targets for a specific condition get to keep a portion of the savings, and they reimburse Medicare for part of the difference when costs are exceeded.
The present study focused on 2013-2015 Medicare claims for the five medical conditions that account for two-thirds of patients enrolled in medical bundles: congestive heart failure, pneumonia, chronic obstructive pulmonary disease, sepsis, and acute myocardial infarction.
Mean baseline payments per episode for those conditions were $24,280 before participation in the BPCI. After hospitals joined, their average payments per episode were $23,993 (P = .41). For a set of matched control hospitals, payments were a mean of $23,901 at baseline and $23,503 in the corresponding follow-up period (P = .08).
That amounted to a $286 payment reduction for BPCI hospitals and a $398 reduction for controls, a difference of $112 (P = .79), the study investigators reported.
Changes in length of stay, readmissions, emergency department use, and clinical complexity of cases from baseline to follow-up periods was not significantly different between BPCI and control hospitals. For example, 90-day mortality increases were seen in both groups, and the degree of increase was not statistically different between the groups.
Those data help fill a gap in research on the BPCI program and BPCI Advanced, a related version of the demonstration project that will have its first cohort of participants starting Oct. 1, 2018.
“Despite the importance of episode-based payment, there has been little research examining its efficacy or determining whether it has unintended consequences, such as hospitals’ selecting patients with relatively less complex conditions to reduce costs and improve outcomes,” Dr. Joynt Maddox and her colleagues cautioned.
It’s unclear why the previous joint replacement study showed a successful reduction in costs under BPCI, while the new study did not. However, patients in the new analysis of the most common bundled conditions were older and had higher rates of poverty and disability.
“As a result of these complexities, patients admitted for medical conditions may have had post-acute care needs that were less amenable to intervention,” Dr. Joynt Maddox said.
The investigators added that hospitals’ lack of effective influence on post–acute-care services may blunt their ability to achieve greater savings under BPCI. Better relationships with skilled nursing facilities, long-term care hospitals, home health agencies, and inpatient rehabilitation facilities could make a difference.
The Commonwealth Fund supported the study. One study author reported personal fees from HHS outside the submitted work, and another reported that he is an associate editor for the New England Journal of Medicine. No other disclosures were reported.
SOURCE: Joynt Maddox KE et al. N Engl J Med. 2018 Jul 19;379(3):260-9.