by Jonathan Pearce, CPA, FHFMA and Darcie Hurteau
With the recent announcement of the next generation of Medicare bundled payment programs (“BPCI Advanced”), many different organizations are filling the internet with opinions about which types of providers should participate in which type of episodes. A frequent topic in this conversation is the “ownership” of the episodes; i.e., which provider should be financially responsible for the surpluses and deficits in these episodes. In the Medicare Bundled Payment for Care Improvement (BPCI) and Comprehensive Care for Joint Replacement (CJR) bundled payment programs, as well as BPCI Advanced, the “participant” receives 100% of the surplus and is responsible for 100% of the deficits, so the ability to create “savings” can have significant financial implications to the participant, and both hospital and physician groups have vied for the opportunity to “own” the episode. Physician ownership was allowed in the initial BPCI program, but in the CJR program and the recently-cancelled EPM program only hospitals could be participants. However in the BPCI Advanced program physicians will have the opportunity to “own” episodes, allowing them meet the requirements in MACRA.
Recently some pundits have opined that physicians should own BPCI Advanced episodes, citing their greater influence on patient care and ability to make positive changes in care protocols than hospitals. This is certainly true, but most physicians know that providing better patient care doesn’t always lead to greater financial rewards, and bundled payment programs exemplify that situation. While physician ownership of major joint replacement episodes (MJR) has often been successful in the BPCI program for reasons described below, the financial results of other types of episodes are far more variable, with the means of creating success largely outside of the control of physicians. In addition, the opportunities for success are much more limited in certain episode types, and the high amount of cost variation combined with low episode volume can create significant amounts of financial risk for the episode owner. While some large physician groups may have the capital to withstand ongoing financial losses, most community physicians don’t want to risk amounts many times higher than their normal compensation over the potential for losses that they can’t control.
How financial results are determined in bundled payment
In Medicare bundled payment episodes, financial results are determined by the relationship between the Medicare provider payments (the “episode cost”) and the target amount. The episode cost is the sum of all Medicare payments to all providers who provide services to the patient within the period of the episode, which is usually the initial hospital stay and 90 days after the discharge. In the BPCI Advanced program targets are computed from historical episode costs. This means financial success is dependent on the participant’s ability to reduce Medicare costs below their historical levels.
An additional factor in target-setting is the “CMS discount”, which is a percentage by which the target is decreased to allow CMS to create savings from each episode. In BPCI these percentages are 2% or 3% based on episode length (90 or 30 days respectively), or 1.5% to 3% based on the participant’s quality scores in the CJR program; however in BPCI Advanced the CMS discount is 3%. Since these percentages are applied to the average historical episode cost, participants must decrease their episode cost more than the discount amount before any savings accrue.
How winners and losers are created in bundled payment
The total cost of a BPCI Advanced episode is the sum of all Medicare payments to all providers from the initial (“Index”) hospital admission through 90 days after the hospital discharge. Knowing that index hospital’s Medicare DRG payments are a fixed amount, savings are generally created in two ways – by reducing or substituting lower-cost care for post-acute institutional services (SNF and IRF), or by reducing readmissions. The extent to which these opportunities exist varies by the clinical category of the episode. For example, readmissions in major joint replacement episodes are generally low (less than 10% of episodes have readmissions) and may be largely unavoidable, so the primary opportunity lies in reducing unnecessary SNF and IRF services. By contrast, the primary opportunity for cost reduction in chronic disease episodes such as CHF or COPD lies in reducing readmissions, since about half of these episodes include at least one readmission. While post-acute institutional care can generally be managed by creating and implementing care management protocols, reducing readmissions is a highly physician-intensive process with fewer guaranteed pathways for success. Therefore, the opportunities for cost savings vary widely based on the clinical definition of episode, and therefore opportunities for physicians in different specialties are also highly variable. Cardiologists will not have the same opportunity to create cost savings as orthopedists do because of the significantly different structure of their episodes.
Assessing opportunity in bundled payment episodes
The differences in these opportunities can be visualized in the graphs shown below. In these graphs, each column represents an individual patient episode, the height of the column is the cost of the episode, and the colors indicate the components of that cost. Because inpatient acute care is paid on a DRG basis, the blue components of these bars that indicate the cost of the "index admission" are generally the same height, with differences being attributable to different DRGs. The purple color above the cost of the index admission generally represents physician services, which rarely represent an opportunity for savings. Therefore, the predominant savings opportunities occur in the yellow, green, and red areas representing home health, SNF, and readmissions.
The graph below shows the episode composition for major joint replacement episodes. It's evident from this graph that about half of patients received SNF services, and relatively few have readmissions. Consequently, the primary savings opportunity is in reducing the number of patients sent to SNF after joint surgery, as well as working with the SNFs to create a shorter length of stay.
The above graph contrasts with the following graph for congestive heart failure. The predominance of red in these episodes shows that readmissions are a major cost driver for CHF patients, with SNF services also adding a significant amount of cost for the higher cost patients. This comparison demonstrates that different strategies and care management processes are necessary to achieve cost savings with CHF episodes than with major joint replacement episodes. This distribution of cost drivers is also found in most medical episodes (sepsis, stroke, COPD, etc.).
The final graph below depicts the episode composition for coronary artery bypass graft (CABG) episodes. Notable in this graph is the predominance of the index admission and physician cost (blue and purple respectively), compared to the relatively small amount of post-acute cost. In these episodes, there are relatively few opportunities to achieve meaningful cost savings, as compared to the opportunities in the episodes discussed above. Because CMS applies the discount to the total episode cost (including the index admission and physician cost) when developing targets, creating savings that exceed the discount is extremely difficult because of the relatively small proportion of post-acute cost in these types of episodes.
Understanding these differences in cost composition is critical for providers in evaluating their potential risks and opportunities associated with "owning" these types of bundled payment episodes. Specialists evaluating BPCI Advanced participation cannot generalize their opportunities from the results of other specialties – they must evaluate the episodes that will occur within their own specialties and assess their own opportunity to create cost savings within the episodes that will be attributed to them.
Opportunity variation among episode types
Another critical factor in physician ownership of the episodes is the extreme amount of variability in cost among episodes, and over time. The graph below shows the average quarterly episode cost for CHF episodes for a number of academic medical centers participating in these episodes in the BPCI program. Each group of graphs represents one hospital participant, with the blue, orange, green, and yellow bars indicating the average episode cost for that participant for each of four quarters in the year.
The hospital leftmost on the graph achieved savings of $4,000 per episode in the first quarter, but by the third quarter it lost about $2,000 per episode. At a volume of 150-200 episodes per quarter, losses of $300,000 – $400,000 would occur in each of those quarters. The majority of the remaining hospitals had similar experiences with only one experiencing consistent surpluses across all four quarters.
This extreme amount of risk variation is difficult for a large medical center to assume, but would be even more significant for an independent cardiology practice at risk for this type of an episode. Physicians who are paid several thousand dollars for treating a patient admitted with congestive heart failure become financially responsible for the costs of post-acute care that may be many multiples of that amount.
Conclusion
Our objective in this paper is not to suggest that physicians should never participate as risk-bearers in bundled payment episodes; in fact, many orthopedists have had significant financial success in owning the major joint replacement episodes in the BPCI program. Those successes have originated from the characteristics of that episode type; namely that there is significant post-acute care utilization in these episodes, that changes in post-acute treatment patterns can be applied broadly across a large number of episodes, and problematic readmissions are few. This combination creates a large opportunity for cost reduction.
Cardiologists, hospitalists and other medical specialists aspiring to duplicate this performance have a significantly different landscape, as shown above. The greatest opportunity with CHF episodes is readmission reduction, which is extremely resource-consumptive, while in CABG episodes there is little post-acute care and consequently a narrow opportunity for success. These difficulties are present in many other episode types; in fact major joint replacement episodes may be the only low hanging fruit for physicians in the entire bundled payment program.
A better opportunity for many physicians is to contract with the participating hospital as part of the "gainsharing" program in which the hospital can share a specified amount of savings or loss with certain physicians. While the upside of these arrangements may not be as lucrative as ownership of the episode, there is rarely any downside risk for physicians in gainsharing arrangements. The BPCI Advanced program allows physicians to be paid as much as 150% of their Medicare payment throughout each episode, which exceeds the payment rates of many commercial payers. Physicians interested in participating in the financial results of Medicare bundled payment programs may find that participating in a gainsharing agreement with the hospital makes more financial sense than becoming a participant on their own.
Jonathan Pearce is the Founding Principal of Singletrack Analytics. Darcie Hurteau is a Senior Director at DataGen.