Limits on savings from bundled payments identified in new research

• Multiyear savings are occurring in bundled payment but at lower rates than seen in a single year.
• The identified savings from hip and knee replacement bundles could allow those bundles to be spread across Medicare.
• Savings from bundled payments for other conditions may require design changes and more time.

As Medicare shifts increasing amounts of its spending into bundled payments, researchers are warning about dwindling savings and are calling for reforms to the programs.

Two studies published this week in Health Affairs examined savings from Medicare’s bundled payment programs: Acute Care Episode (ACE), Bundled Payments for Care Improvement (BPCI) and Comprehensive Care for Joint Replacement (CJR). The BPCI program has been replaced by the similar and ongoing BPCI Advanced (BPCI-A) model, the second cohort for which launched Jan. 1.

The programs use provider bonuses and payment penalties, based on providers’ performance against their previously benchmarked results for Medicare spending and on quality performance on targeted treatments. ACE and BPCI were voluntary, as is BPCI-A, while CJR has been mandatory for most participating hospitals.

Findings from the examination of the first three years of BPCI, as detailed in the first study, included:

• Spending was reduced by 1.6% over three years for lower-extremity joint replacement (LEJR) — the highest-volume episode.
• Savings were driven by early participants.
• Patient selection drove 27% of episode savings.

Why the findings matter

Amol Navathe, MD, PhD, an assistant professor of health policy and medicine at the University of Pennsylvania and co-author of both studies, said the savings driven by patient selection “are not real cost savings.”

Instead, the findings bore out concerns that providers in bundled payment models would pick “favorable patients to beat the benchmark rather than truly improve quality,” Navathe said in an interview.

Another significant finding was the possibility that providers joining BPCI-A in 2020, following the first cohort that started in late 2018, would have lower savings than already-participating providers, which in turn had lower savings than the 3.9% that CMS found in the first year of BPCI.

The divergent results, Navathe said, may stem from providers with less expertise in bundles moving into such models in response to the recently revised Medicare physician payment system, which favors providers participating in advanced alternative payment models.

Implications across bundles

The second study compared previously published results from ACE, BPCI and CJR and concluded that bundled payment maintains or improves quality while lowering costs only for LEJR (i.e., hip and knee joint replacement surgery).

It may take longer for savings to accrue for other conditions because they generally involve sicker patients, Navathe said. But he was optimistic savings eventually could appear in bundles for other conditions.

The findings led the authors to urge rule changes in bundled payments, to account for patient-level heterogeneity and incorporate risk stratification for specific conditions.

Other findings of the second study included:

• Quality was the same or better across all episodes.
• Evidence supports expanding LEJR bundled payment throughout Medicare.
• There is a risk of provider reluctance to accept patients with certain conditions that are not associated with savings.
• Program changes may be needed to produce better value from caring for patients with non-LEJR conditions.

“Newer models need to be more refined around risk adjustment, particularly given patient-level heterogeneity,” Navathe said.

The findings may be significant for BPCI-A, which in 2020 has added bundled payment episodes for seizures, inflammatory bowel disease, bariatric surgery and transcatheter aortic valve replacement, bringing the total number of bundles to 31. Major joint replacement of the lower extremity, the most popular episode selected during the 2018 enrollment period, has been expanded to include outpatient knee replacements.

“That’s great that we’re seeing Medicare test more innovative models,” Navathe said.

The findings also could carry significance for the CJR model, which the Trump administration is considering extending for three years and revising.

The findings also may influence the number and design of offerings in the commercial market, where a small (2% in 2017) but growing share of payments comes through risk-based bundles, according to recently released data from the National Alliance for Healthcare Purchaser Coalitions.