Which Episodes Should I Bundle for BPCI

by Jonathan Pearce, CPA, FHFMA and Coleen Kivlahan, MD, MSPH

Applicants to the Medicare Bundled Payment for Care Improvement (BPCI) program will be busy spending December identifying episodes for which they should participate when they "go live" in April. Several episode families are common candidates for participation, and an overview of their characteristics is described below.

Major Joint Replacement: Participants really need a reason not to participate in bundled payments with major joint replacements. These DRGs have most of the characteristics necessary for success, including:

  • Significant volume – The major joint replacement DRGs have some of the highest numbers of Medicare admissions of any group of DRGs. High episode volume reduces random variation in episode costs and creates more consistent, measurable results, and  also creates a significant opportunity for financial return.
  • Past history of success – Major joint procedures were included in the CMS Acute Care Episodes (ACE) demonstration project that preceded the BPCI program. Most of those participants were able to achieve significant savings through care redesign and internal cost savings, and those results are available from CMS website.
  • Major joint episodes are composed primarily of the index admission and care from post-acute providers, creating an opportunity to develop cooperative care pathways with those post-acute providers. Readmissions are infrequent relative to their occurrence in other episode families.
  • The vast majority (about 75%) of post-discharge cost occurs within 30 days after discharge. This increases the manageability of this care. It also allows participating in 90 day episodes (which require a 2% discount to CMS rather than the 3% discount required of shorter episodes) with little additional risk, since utilization during the last 60 days of the episode is relatively small.
  • The incidence of high-cost cases is significantly lower than other episode families. Most patients electing joint replacement are healthy overall. This reduces the risk of loss compared with other episode families.

One clinical and financial risk in this episode family is the occurrence of unexpected and non-elective hip fractures. Hip fracture cases have a significantly higher incidence of readmissions, and a higher disease burden, post-acute cost and total episode cost. A growing proportion of fracture cases within a low volume population can cause the average episode cost to increase significantly, as has occurred with several of our clients. This can occur due to physician specialization, Level 1 trauma facilities, and transfers; therefore it is essential to evaluate fracture incidence in your population, develop a specific management plan for these patients and assess their effect on future average episode costs.

Congestive heart failure: Participants frequently choose this episode family as an opportunity to participate in medical DRGs. These episode families create a greater challenge, but potentially more opportunity if those challenges can be met. CHF episodes have a significantly larger amount of episode cost variation than surgical episodes, and a much larger proportion of post-discharge cost is in readmissions rather than care from post-acute providers. Therefore, participants must have a strong readmission prevention strategy focused on readmissions for CHF and related comorbidities (pneumonia, arrhythmias, COPD, etc.). Different from major joint patients, CHF patient care is almost evenly distributed throughout the 90 day post-acute period, which means that care management techniques must continue to be effective two to three months after discharge.  The shorter 30 day episodes may be more appropriate for participants who believe they can efficiently manage care provided within a short period after discharge, but are less confident of their ability to manage it over longer post-discharge periods. In addition, these episodes have lower volume than surgical procedures, which creates a larger quarterly episode cost variation. This results in significant swings in quarterly reconciliation amounts, which means that participants in these DRGs must evaluate results over a significantly longer period of time than is necessary for other episode families having higher volumes. Short term (i.e., quarterly) results are likely to be inconclusive because of high rates of episode cost variability.

COPD and pneumonia: This episode family has similar characteristics to CHF episodes, although with a lower incidence of high-cost cases. Readmissions are a primary cost driver, and their incidence occurs almost evenly throughout the post-discharge period.

Cardiac valves: These DRGs provide less opportunity to decrease Medicare cost, simply because the majority of the cost occurs during the index admission (in which the hospital was paid a fixed DRG payment), rather than during the post discharge period. While 75% of the cost in CHF episodes occurs in the post-discharge period, for cardiac valve episodes about 25% of the cost occurs in the post-discharge period, with the remaining 75% occurring during the index admission. For most health systems, the primary strategy for cardiac valve episode participants is to reduce internal hospital costs, rather than improve care management during the post-discharge period. While some participants have had success in this episode family, savings opportunities from post-discharge cost reductions appear to be lower than in other families.

Percutaneous coronary intervention (PCI): PCI episodes offer an opportunity to participate in surgical DRGs which have similar characteristics to CHF and COPD medical episodes. Slightly less than half of the episode cost is in the post-discharge period, providing less opportunity to create cost savings than in major joint procedures, and there is a significantly higher readmission rate  than occurs in major joint procedures. Therefore, participants in this DRG must develop clinical strategies for both post-acute provider and readmission cost management.

One factor to consider is that in this episode family, there is a clinical direction to manage low risk patients in a one day or observation stay, yet there is an increasing rate of denial of one day stay episodes. Since one-day stays are generally less costly than longer stays, eliminating those episodes will increase the clinical risk of the hospitalized population, resulting in higher average episode costs of the remaining episodes. While this change may be accommodated in the trend factors on an ongoing basis, the inclusion of one-day stays in the base period will reduce the initial target prices below those that may be achievable during the performance period. In addition, the higher risk PCI patients may also require intensive additional procedures during their 90 day episode resulting in higher total cost and more outliers. 

Stroke: Costs in the stroke episode family are almost equally driven by the index admission cost, SNF costs and inpatient rehabilitation costs, which together account for about 80% of the episode cost. Readmission costs account for less than 20% of the cost in this episode family, so the primary strategy is in properly managing the post-acute care continuum. Generally more than 2/3 of the post-acute cost occurs in the first 30 days after discharge, so the exposure to longer episode lengths is less significant than in CHF and other medical episode families.

Episode participation analysis: The above overviews of several popular episode families will hopefully provide a starting point for potential participants. However, additional analysis is always necessary to identify the specific characteristics of each episode family in a participating institution. This process, and the analytical tools used in it, are described in BPCI Episode Selection – An Analytic Approach. Other helpful articles describe the process for evaluating financial risk, and in understanding how random variation and “luck” can influence results. Using these tools, participants can understand the specific characteristics of their own episode costs, make participation decisions, and formulate strategies for success.

Jonathan Pearce is a Principal with Singletrack Analytics. Dr. Kivlahan is a family medicine physician and a Senior Director at the Association of American Medical Colleges, where she directs their BPCI convener program.

Which Episodes Should I Bundle.pdf310.34 KB