By Steve Wiggins, Founder and Executive Chairman
SUMMARY:
Medicare’s Bundled Payment for Care Improvement (BPCI) initiative is showing powerful evidence to support widespread expansion of bundled payments in the Medicare fee for service program. Analysis of results from approximately 1,000 Participating health care organizations demonstrates savings to Medicare of $200 million from these early entrants. Savings to Medicare from this group of Participants in the BPCI initiative are currently running at $20 million per quarter.
All Participants in this large cohort show steady improvement over time, providing evidence that a wide range of organization types can succeed in bundled payments. Physician groups, however, achieve the highest savings rates for acute-triggered episodes and reach those savings levels faster than hospitals. Additionally, the 1,000 organizations in this cohort have demonstrated success in both surgical and medical episodes of care. Four in five patients who begin episodes at these 1,000 organizations do so because of an emergency admission to the hospital. This provides encouraging proof of the broad applicability of bundled payments across episode types and dispels the prediction that bundled payments may only succeed with planned surgeries.
Results from Medicare’s voluntary Bundled Payment program reveal substantial reductions in the number of patients discharged to post-acute facilities, a reduction in the length of stay at those facilities, and fewer patient readmissions following a hospitalization. These utilization results were achieved without a rise in unadjusted mortality among this challenging patient population. Because they are Medicare fee for service patients with an inpatient stay, BPCI patients often have substantial co-morbidities giving rise to a high level of complications during an episode.
These results provide strong support for Medicare’s continuation of the bundled payment program and supports expansion into additional diseases and conditions. They also suggest that population health programs, such as ACOs, would benefit by incorporating bundled payments into their strategies for engaging specialists, as well as for lowering costs and improving outcomes.
Detailed Results:
Estimated Savings to Medicare:
Table 1.
Table 1 displays the net savings, by quarter, for the Medicare Trust Fund. Due to data challenges described below, the savings to Medicare were lower for some quarters than what should be expected in the future. Nevertheless, Medicare saved over $156 Million during the early periods of the BPCI Demonstration from this cohort of Participants.
The CMS third party vendors managing the BPCI initiative encountered numerous challenges related to the underlying data used to establish baseline prices. Soon after launching the program, CMS and its vendors realized that baseline prices had been calculated with a large swath of missing claims. This led to a program-wide ‘re-pricing’ of the Target Prices for every episode. Participants were treated financially in the same manner as Track 1 MSSP ACOs for two quarters (Q1 and Q2 2014), with no financial penalty for those organizations without savings sufficient to pay Medicare the guaranteed discount during those quarters. CMS allowed these early participants to review the new benchmark data and determine for which episodes they wished to voluntarily accept full downside risk in subsequent quarters. Those participants, almost exclusively hospitals, moved to full risk in the third quarter of 2014.
Physician Groups were a very small part of the initial enrollment into BPCI. However, CMS had ‘opened the window’ for new entrants during 2014 and a substantial number of physician groups entered the BPCI initiative, with the vast majority of entrants ‘going live’ with full risk programs in the last half of 2015. Unfortunately, the bundled payments program revealed material inaccuracies in the Medicare data base that is used to link physicians to physician groups and to link physician groups to patients. The Internet-based Provider Enrollment, Chain and Ownership System (PECOS) had led to a myriad of inaccuracies in the initial quarterly reconciliations of physician group Participants.
CMMI undertook a complex process to correct the PECOS data base and those corrections effectively led, once again, to a re-pricing of the entire physician group cohort of the BPCI demonstration. CMMI followed its precedent of not punishing Participants for pricing inaccuracies and treated physician groups in the same manner they had treated hospitals, using the Track 1 MSSP ACO approach to not holding participants responsible for losses.
Consequently, savings to Medicare from physician group participants were lower for the six quarters that passed before physician groups were allowed to drop episodes and return to full risk. With the benefit of hindsight, it is now clear that despite the re-pricing of episodes in 2014 for hospitals and again in 2016 for physician groups, Medicare still achieved demonstrable savings that dwarf any other value based payment initiative.
BPCI: Demonstrated success lowering spending
Limitations of this study:
This analysis includes information covering 950 health care organizations participating in Medicare’s BPCI initiative. We estimate this represents over 55% of the entire BPCI participant universe and approximately half of the spending managed under Medicare’s BPCI initiative.
Important information about these 950 organizations:
All organizations operate through a Participant Agreement with Remedy Partners, an Awardee Convener approved by Medicare to participate in the Medicare BPCI initiative. Remedy has a Master Participant Agreement with CMS and acts as the ‘general contractor’, handling analytics, software development and implementations, training and administration of all compliance and reconciliation services.
Over 900 organizations, of the 950, have adopted a common software system that is integrated with their electronic medical record software. The software identifies patients, in real time, who are accreting into a BPCI episode and is a workflow, decision support and patient/caregiver engagement platform that was developed specifically for episodes of care. All Participating organizations have participated in the refinement of their shared software platform.
All 950 participants use a data analytics platform that delivers program level, provider level and patient level metrics and reports that have enabled regular refinements to their programs and enabled leadership interventions for fledgling programs.
All programs were supported by $200 million of working capital funding that was provided by the Awardee Convener to assist in the financing and launch of the local bundled payment programs.
Over 80% of Medicare’s BPCI initiative runs through an Awardee Convener, such as Remedy Partners. An Awardee Convener is a Special Purpose Entity (SPE) that performs the role of general contractor, participant recruiter, primary risk-taking entity, network developer and primary administrator of these relatively complicated programs. Most Awardee Conveners share proportionally in the risk and rewards of these programs, with Participating provider organizations eventually taking the lion’s share of the risk/reward.
ACOs are another form of Special Purpose Entity (SPE) in the population health field. Disease management companies, Independent Practice Associations (IPAs), Management Service Organizations (MSOs) and Physician/Hospital Organizations (PHOs) are other types of SPEs common in commercial insurance and Medicare Advantage programs.
Generally, payer organizations have been unsuccessful in achieving meaningful adoption of at-risk contracting in population health or bundled payments, without SPEs to recruit participants, provide capital and expertise.
Observations of the BPCI initiative:
This analysis suggests that Medicare, through CMMI, succeeded in launching and managing a complicated bundled payment program and has demonstrated success in attracting a large group of participants, lowering health care spending overall, achieving meaningful savings to Medicare, and improving health outcomes.
For those considering launching commercial bundled payments, however, Medicare’s initiative is instructive:
Health care providers prefer to avoid assuming downside risk. Most health care organizations already operate with low margins and many lack the balance sheet stability required to assume risk. Special Purpose Entities are likely necessary to achieve meaningful adoption of bundled payments and to overcome the reluctance, lack of capital and lack of expertise in these complex programs. Awardee Conveners are the Special Purpose entities that drove most of the enrollment in BPCI, garnering over 62% of Participants and a higher share of spending. Awardee Conveners are analogous to ACOs, IPAs, PHOs or other entities in health care who provide the capital, administrative capability, software and analytics to support successful programs. In the case of the cohort studied in this analysis, over $200 Million was invested by the Awardee Convener in the working capital for these programs.
BPCI ‘triggers’ an episode on the basis of a patient’s acute care admission to a hospital. This is due to the limitations of administering a program through 17 different payment contractors. For single payer initiatives, far greater savings and more impactful programs will be achieved by triggering the opening of an episode of care ‘at diagnosis’. Triggering an episode at diagnosis will capture the most important decision points and drive patients and physicians to carefully reconsider many habits that are out of synch with patient need, best practice or greatest efficiency.
BPCI encountered massive challenges with fee for service data that was never collected with the intent of driving episode pricing. This data challenge will be exacerbated in commercial programs where the girth of Medicare’s volumes won’t allow history to guide how prices are set. Combining a wide range of data sources will be required, and new applications of math theorems and deep machine learning will likely improve the accuracy of Target Prices of episodes of care.
Established payment arrangements can conspire against efforts to achieve greater efficiency. The prospective payment methodologies for hospitals (Diagnosis Related Groups) and home health care services (Home Health Resource Groups), used by Medicare, have led to distortions in the incentives they create. Keeping a patient in a hospital a bit longer can drive much lower post-acute costs. Careful titration of home health services and freedom to use these services in ways currently prohibited by CMS would likely unlock further improvements and savings. Likewise, allowing at-risk providers to control other payment arrangements would lead to greater innovation in the use of remote monitoring, telehealth and in-home advance services.
Many bundled payment experts believe that population health programs, such as Accountable Care Organizations, should use bundled payments as the primary vehicle for managing expensive episodes of care. The Medicare performance evidence from the ACO and BPCI initiative would support that view. Proponents view population health strategies as particularly valuable for engaging primary care physicians in the identification of high risk beneficiaries and undertaking strategies to promote health maintenance and disease prevention. Bundled payment programs, on the other hand, are proving to achieve lower costs and favorable outcomes for patients who may have failed this prevention path and require expensive planned or immediate therapeutic interventions. The complexity of cases served by BPCI also supports the suggestion of Michael Porter, Harold Miller and others that bundles may also be appropriate for Chronic Conditions.
Recent studies have shown that ACOs are struggling to generate sustainable savings to share with Medicare.[1] History and experience reinforce these concerns. Widespread failures during the 1990s left many hospitals and physician groups reeling from their initial attempts to accept population-level risk.[2] Managing risk at the population level requires a vast array of cost control strategies and few provider organizations have developed the required expertise to control both ‘incidence’ and ‘performance’ risk. That is one reason that ACOs are struggling to achieve their potential.
To make further progress, ACOs and their sponsoring organizations will need to adopt sophisticated strategies that tackle all aspects of medical spending. More particularly, as these organizations seek to include specialists and hospital-based clinicians in care redesign, bundled payments will prove especially useful.
Conclusions:
The information disclosed in this Brief provides strong evidence that bundled payments should be a cornerstone of Medicare payment policies. Indeed, they perform better than any other initiative in the history of the Medicare fee for service program. By CBO’s own estimates, Medicare stands to save $46 billion by aggressively adopting this payment model. BPCI has also attracted the largest universe of Participant health care organizations in a payment demonstration, revealing intrinsic acceptance of this approach to organizing and financing health care around a patient’s episode of care.
Population health programs would likely benefit from use of bundled payment strategies and tighter integration between the two complimentary payment models.
Implications for Non-Government Programs
Commercial insurers should piggyback on Medicare’s expanding bundled payment initiative to drive greater use of this payment model for their employer and government programs. For commercial insurance programs, bundles are a better than population health and encourage greater competition among health care providers. They also empower a wider range of organizations to participate and drive innovation.
Commercial insurers also have the ability to overcome the limitations of the Medicare initiative by triggering episodes at diagnosis instead of at acute admission. Commercial plans also have the flexibility to incorporate bundled payments and patient steerage into how they administer prior authorization and design benefit plans. Incorporating bundled payment arrangements into health benefit plan designs holds the potential to accelerate adoption and innovation.
Surprisingly, the federal government has for now assumed the role of the leading innovator in health care payment reform. This is a powerful incentive for private insurers and self-funded employers to take advantage of the momentum created by this payment demonstration.
1 McWilliams JM, Chernew ME, & Landon BE. Medicare ACO Program Savings Not Tied to Preventable Hospitalizations Or Concentrated Among High-Risk Patients. Health Affairs 2017; 36(12):2085-2093.
2 Burns LR & Pauly MV. Accountable Care Organziations May Have Difficulty Avoiding The Failures Of Integrated Delivery Networks of the 1990s. Health Affairs 2012;31(11):2407-2416.
“The statements contained in this document are solely those of the authors and do not necessarily reflect the views or policies of CMS. The authors assume responsibility for the accuracy and completeness of the information contained in this document.”