Background

Until recently, Medicare, Medicaid and most private insurers paid physicians for each individual service provided. The aptly named Fee-For-Service (FFS) approach is considered a cornerstone of our current healthcare system. Critics of this approach argue that providing a fee for each individual service incentivizes increased volume rather than quality of care. Many observers believe that CMS is consciously making efforts to limit the exposure of the system to FFS.

In January 2015 the Secretary of Health and Human Services, Ms Sylvia Burwell, published a roadmap in the New England Journal of Medicine for transitioning to alternative payment models and value-based payments for the Medicare program.2 We have provided a detailed analysis of Burwell's approach in a previous paper.3 In her article she articulated three overarching themes. The first two have bearing on this review:

1. Use incentives to motivate higher-value care by increasingly tying payment to value through alternative payment models.

2. Changing the way care is delivered through greater teamwork and integration, more effective coordination of providers across settings, and greater attention by providers to population health.

Accountable care and bundled payments are two separate methods designed to achieve cost saving and improve quality. Accountable Care Organizations (ACOs) are designed to have a healthcare delivery system (the specific elements of which might vary) take responsibility for a population of patients and, as the name suggests, become accountable for the care of that patient group. The concept is that, in the FFS environment, there is little incentive for hospitals or physicians to invest in infrastructure that does not result in something that is directly reimbursable or that allows care redesign. Moreover, if a single entity is responsible for a given population, that entity is vested in coordinating care and being efficient in delivering it. ACOs profit by keeping patients healthy and by promoting efficiencies when they are sick.

Payment bundling is a related but distinct initiative. It is a related concept in that it too is anticipated to produce cost savings by converting from a FFS to a fee-for-episode or fee-for-results environment. The BPCI initiative seeks to offer providers a combined payment for a single illness or a single course of treatment. A single course of diagnosis and/or treatment is sometimes termed an ‘episode of care’. The explicit assumption is that, by offering a bundled payment, providers will police themselves and become jointly accountable for that episode of care. As in the case of ACOs, if providers are efficient they can improve their own bottom line by being careful in how they manage resources during that episode of care. Payment bundling and ACOs are both examples of ‘at risk’ payment initiatives where the providers assume a greater degree of risk for cost overruns than in the typical FFS environment.

It is important to distinguish this form of payment bundling from the current procedural terminology (CPT) code bundling that has become familiar to readers in the past few years. CPT code bundling refers to the redefining of multiple CPT codes for services likely reported together into a single CPT code for this combination of services.4 Code bundling can refer to the combination of surgical codes and supervision and interpretation (‘S and I’) codes into a single procedural code, or to the combination of related S and I codes into a single code. In practice, both code bundling and procedure (or payment) bundling can be expected to reduce overall payment for a given set of services.4 ,5

The initiative

Bundled payments seek to align incentives for providers. In this case, ‘providers’ represents a broad term and includes physicians, non-physician providers, hospitals, and other facilities. Indeed, bundled payment models are under active development. CMS announced which healthcare organizations are participating on 31 January 2013.6

There are four models currently being considered as part of the BPCI initiative:

• The first focuses on acute care inpatient hospitalization with a standard discount to Medicare from routine Part A inpatient payments.

• The second and third models involve a retrospective bundled payment arrangement where actual expenditures are reconciled against a target price for an episode of care.

• The fourth model involves a prospective bundled payment arrangement where a lump sum payment is made to providers for the entire episode of care.7

Looking at these models in greater detail, one can envisage different direct impacts on physician providers. In the first model the episode of care is, by definition, the inpatient stay in the acute care hospital. CMS will continue to pay physicians separately for their services under the FFS model using the Medicare Physician Fee Schedule. In the second model the bundled payment will include the inpatient stay in the acute care hospital and all associated physician and non-physician services during the episode. The episode ends up to 90 days after hospital discharge and thus will include services provided during the post-acute care period. The third model is in some ways less ambitious than the second since it only involves the post-acute care period, although one requires a stay in an acute care hospital to be a part of it. This period must begin within 30 days of discharge from an acute care inpatient stay and can extend up to 90 days after the initiation of the episode. The fourth model is perhaps the most ambitious of all of the BPCI efforts. Providers will receive a single, prospectively designated payment for the episode of care. This payment will be made to the hospital that provides the acute care and is intended to cover all services that occur during that single stay. To be clear, physicians will submit ‘no-pay’ claims to CMS for the services they provide but will receive no separate payments. Rather, they will be paid out of the prospectively determined bundled payment. Finally, readmissions that are felt to be related to the original rationale for admission and occur within the first 30 days will need to be paid for from the original bundle.7

BCPI and the neurointerventionalist

There are 48 different care ‘episodes’ that are available for models 2–4 in 2015. The one that impacts neurointerventionalists the most is stroke, which includes multiple DRGs (061–066) for ischemic and hemorrhagic infarction. The authors encourage readers to consider the implication of each of the four bundled payment models going forward. Potential mechanisms of payment could include bundling of all imaging, emergency services, catheter intervention, critical care, and inpatient costs. Depending on the model, one could expect post-acute care discharge recovery in a rehabilitation facility to be included. Who will determine appropriate distribution of payments for professional services? Is it the hospital, a multidisciplinary team of physicians, or possibly both? Is payment determined by prevailing Medicare Part B rates or by some new formula? Critically, the reader should realize that these approaches are being tested currently.3 ,7

Early results for BPCI

The early results from BPCI have been inconclusive. Evaluations of models 3 and 4 were limited by the small numbers of national participants precluding any reliable conclusions. Model 2 had mixed results. Across all model 2 episodes, the share of BPCI patients discharged to higher cost skilled nursing facilities was five percentage points lower in the first intervention quarter than in the baseline period. As a correlate, the patients discharged to home health care, which is typically a lower cost provider, increased.

Readmissions declined 2.5% more for BPCI participants than for the comparison providers when evaluating the 30 days after discharge from the anchor hospitalization.

One early potential source for concern about quality of care originated from data collected on surgical orthopedic care excluding spine episodes under model 2. Emergency department use for BPCI patients in these episodes increased. This begs the question about financial incentives motivating early discharge or cheaper less effective post-acute care leading to elevated downstream costs not from readmissions but from increased use of the emergency department or more frequent outpatient visits. Further data are being collected to answer these questions.6