I have been involved with Bundled Payments for Care Improvement (BPCI) Model 3 from both the post-acute and acute care side of the program. This vantage point has afforded me the opportunity to see this model has some issues that strike at the heart of making bundled payments work operationally and financially for post-acute providers.
Rationale for participating now
The argument I have heard most frequently for post-acute care (PAC) providers jumping into BPCI is that payment models like this are the wave of the future. Providers that are early adopters will be rewarded with “ahead of the curve” insights into how these pricing mechanisms work. Ultimately, those post-acute providers taking the plunge now will be viewed more favorably by other healthcare system actors.
These are both reasonable arguments. However, there are three practical issues that I believe call into question the foundational underpinnings of BPCI from a structural standpoint as well as from where PAC providers stand in the healthcare delivery continuum.
- BPCI Model 3 is, in its essence, a pricing experiment with no change in how real-time Medicare fee for service (FFS) reimbursement to providers is structured. BPCI uses a post payment reconciliation process that can occur between six and nine months after services have been rendered. There is a built-in disconnect between clinical decisions that are made, both in the PAC facility as well as in the community setting, and the eventual financial outcomes calculated by CMS off of claims data. This delay between services being provided and financial results being reported makes efforts to address system and people problems related to negative outcomes a tortured proposition at best.
- Bundled payment initiatives are predicated on controlling the health and healthcare delivery of sick older adults over periods of 30 to 90 days. As patients move through the healthcare delivery continuum, the PAC providers’ ability to control outcomes becomes more and more tenuous—regardless of the lofty rhetoric of integrated networks that is tossed around. This loss of control is complicated by the social and environmental situation in the community where patients are discharged and further complicated by their daily personal choices. To think that success with a CHF bundle, for example, could be undermined by an ill-timed $2.49 bag of potato chips is rather sobering.
- BPCI Model 3 hinges on the PAC provider being able to gain knowledge of the hospital generated final billing diagnosis-related group (DRG). The DRG is used to determine if a patient admitted to a PAC facility falls into one of the categories chosen by the participating PAC provider.
Kindred Healthcare's experience was referenced in a Jan. 19 American Hospital Association report on Medicare’s Bundled Payment Initiatives. It was noted:
“For less complicated patients (as opposed to long-term acute care patients), it has had difficulty determining which patients are in BPCI since hospitals do not know the DRG assignment when the patient is discharged to a post-acute facility.”
Since PAC providers do not know with certainty what patients fall within their chosen DRGs, they are left to make an educated guess or rely on newly-created algorithms that attempt to determine final billing DRG assigned by the hospital coder. This is an inexact science at best and a significant foundational weakness of BPCI Model 3.
The following is an excerpt from the annual report on BPCI the Lewin Group prepared for the U.S. Centers for Medicare & Medicaid Services (CMS):
“Reducing (skilled nursing facility) length of stay directly reduces SNF Medicare revenues, which could make this a difficult trade-off for SNF EIs (episode initiators). They would need to calculate Final CMS BPCI Models 2-4: Year 2 Evaluation and Monitoring Annual Report 18 their potential financial rewards under BPCI for reducing length of stay in comparison to foregone daily Medicare payments.”
It is ironic that a study commissioned and paid for by CMS would recommend providers undertake this analysis that could well result in providers opting out of BPCI.
I recently attended a seminar that touched on bundled payments. The presenter represented a small Midwestern convener with a dozen or so facilities involved in BPCI Model 3. The presenter said the facilities reduced their length of stay (LOS) for patients in the bundle from 30 to 16 days.
The presenter said the average facility received a Net Payment Reconciliation Amount (NPRA) of $1,500 per month, essentially the difference between the CMS established benchmark and the actual cost of care for the length of the bundle. The average facility had four patients with their chosen DRGs per month. The following two scenarios demonstrate the impact of involvement in BPCI with LOS being reduced compared to the traditional Medicare service provision.
Facility in BPCI
16 days LOS X $500 Medicare FFS reimbursement per day X 4 patients + $1,500 = $33,500
Facility in traditional FFS Medicare
30 days LOS X $500 Medicare FFS reimbursement per day X 4 patients = $60,000
The difference between the two approaches is $26,500 and represents the foregone daily Medicare payments referenced in the Lewin Group’s report and experienced by the BPCI participating provider.