Skip to main content

Analyzing CMS’ Responses to Bundled Payment for Care Improvement: Is It Still A Good Deal?

The Centers for Medicare and Medicaid Services’ (CMS) responses to applications for the Bundled Payment for Care Improvement (BPCI) demonstration program have raised more questions than they answered.  While we understand the government’s need to reign in the myriad bundle definitions, the degree to which CMS has gone from “define your own” to non-negotiable specifics raises some significant concerns. CMS has:
  • defined its own episode bundles and exclusions; 
  • created required bundle families; 
  • calculated regional average prices; and 
  • established an outlier methodology. 
CMS is also considering establishing episode prices that are a blend of hospital-specific and regional data, making the demonstration program look much more like a typical prospective payment system.

There is still time to provide feedback to CMS as they continue to develop the methodologies that will apply under the BPCI program.  CMS is accepting questions and comments, via e-mail, and we suggest facilities review the information they receive and understand its implications.  Many awardees have been granted interviews with CMS and they have until November 28 to notify CMS of their continued interest in participating in the program.  CMS expects that the Excel workbooks they provided with their own bundle calculations will provide the necessary information for awardees to make these decisions, and that checking the calculations is unnecessary.  Our best advice: proceed with caution and do your due diligence.

Episode Bundle Definitions and Prices:


Most hospitals expended considerable resources analyzing the claims-level data to identify care patterns, calculate prices, and determine whether they could achieve savings under a discounted price.  CMS has recalculated all episode prices using its own set of standard exclusions and outlier limits that are based on regional averages.  Eventually, CMS will need to trend those prices forward to the program year.  Awardees need to review these calculations for the following:
  • Can you match the CMS number?  
    • We have confirmed that CMS used the entire United States to identify episodes and calculate prices for candidate awardees; and
    • CMS is not currently using all of their stated methodologies in the calculations.
  • Is there still opportunity to achieve savings after all of the exclusions have been applied?
  • Are there additional exclusions that should be applied, such as cancer Diagnosis Related Groups (DRGs)?  CMS has since issued new episode group definitions with some additional exclusions.
  • What factors will CMS use to trend the data forward?  Are they appropriate and complete?
  • How does the application of outliers affect the price and the opportunity to create savings?

Bundle Families:


The original Request for Applications was clear in requiring applicants to bid on all episodes within a Medicare Severity (MS)-DRG family; however, each DRG-based episode within the family must have its own price.  This limitation forced applicants to include episodes that might have smaller volumes.  For example, to include DRG 470 (major joint procedures without major complication or comorbidity (MCC)), applicants were also required to include DRG 469 (major joint replacement with MCC) despite the fact that the latter DRG has much fewer cases. CMS has now broadened the definition of family for some episodes, creating even more small sample sizes - the smaller the sample size, the greater the risk that results will not be based on effective medical care.  For example, the CMS “COPD, bronchitis/asthma” episode family includes the bronchitis/asthma DRGs (202 and 203) in addition to the COPD DRGs (190-192), which increases the possibility of having DRGs with small sample sizes.  This increases the risk of participation in the BPCI demonstration, since the smaller sample sizes have greater variability.

Does the inclusion of additional DRGs for a particular family create too much risk to manage?

Regional Average Prices:


The inclusion of low-volume DRGs in the required families of episodes creates the problem of establishing bundled payment targets based on limited (or non-existent) historical data.  CMS proposes to solve this problem with the introduction of a blended pricing methodology, using regional average costs, or multi-years’ data, or both, combined with hospital-specific costs to increase the accuracy of the episode pricing targets.
The regional average prices are an interesting twist in the process.  CMS had explicitly warned applicants not to look at any data for hospitals other than themselves.  Now that data will be used to set outlier thresholds and “normalize” prices for those DRGs with significant variability.  If CMS chooses to use multiple years of data, how relevant is that older data to current practice patterns and how will blended prices be calculated?  CMS has not made those determinations yet, but they are proposing use of the Empirical Bayes statistical method.  Applicant awardees need to consider:
  • Can the regional averages be replicated/audited?
  • Will the industry have access to additional years’ data if CMS chooses this approach?
  • What is the range of potential prices for an episode, from lowest to highest, given different blend percentages?
  • How will the regional averages be adjusted for different wage indexes?  Will your hospital start out at a disadvantage because of this calculation?
  • How does your hospital-specific price compare to the regional average?  Will you start out at a disadvantage because your facility’s price is higher than others in your region?

The DataGen/Singletrack Team Approach


DataGen and Singletrack Analytics have been working together to closely monitor the proposed BPCI changes, develop models to evaluate the effects of these changes, and provide applicant awardees with the information they need to continue to assess their participation opportunities.  We provided data and analytical reports to more than two dozen applicants, and have worked with a number of clients through the application phase of the BPCI process. From this experience we have developed a thorough understanding of the data and analytics required to drive informed bundled payment participation decisions.  Through our interactive analytic models and consultative support, we keep our clients up-to-date on the effects of the proposed changes and support them through the CMS interview and response process.

According to CMS, the bundling and pricing methodologies are still preliminary and subject to change.  With the proper analytic support, you can help CMS craft a BPCI demonstration program that makes sense and works.  To be an active participant in the process, you need to understand it.   Let our team provide the information you need to confidently evaluate and make a solid business decision.

For more information, contact Gloria Kupferman at gkupferm@hanys.org or at (518) 431-7968.

Popular posts from this blog

What is a Community Health Needs Assessment? 4 Tips to Start

Hospitals typically conduct a  Community Health Needs Assessment  to comply with state requirements or to maintain 501(c)3 status. However, emerging trends around health outcomes and health equity have sparked organizations to update and better align their CHNA processes toward highlighting community needs, equity, population health concerns, service access, affordability and quality. In a research study,  The National Library of Medicine  found that "Social determinants of health impact 80% of health outcomes from acute to chronic disorders, and attempts are underway to provide these data elements to clinicians." Because of the short- and  long-term effects of SDOHs , it's important that hospitals assess community needs. This way, they can find solutions to improve quality of life, identify underserved populations and establish connections with the community. What can your organization do to revamp its CHNA process to focus on community needs, equity, care access, afford

Community Health Assessment Toolkit: Data Collection Methods

Why should you include data collection methods in your Community Health Assessment (CHA) toolkit? A CHA is like an electronic health record for a county, Metropolitan Statistical Area or region. Done well, the CHA captures clinical and social needs, informs options for new service delivery, facilitates collaboration among community stakeholders and ultimately can impact health outcomes.  Public health departments today must collect data on everything from diabetes outcomes to housing, income, immunizations and many other measures. Read on for the top methods for collecting the most challenging yet insightful data.  Community Health Assessment data collection methods  Like an EHR, the CHA includes defined components. The National Association of County and City Health Officials’ Mobilizing for Action through Planning and Partnerships (MAPP 2.0) model has several components and three assessments under the MAPP 2.0 model Mobilizing for Action through Planning and Partnerships Assessments: