CMS’s long-overdue announcement about Medicare’s next generation of bundled payments was like a headline in the New York Post: It tantalized with just enough information to lure you in, but left would-be participants wanting to know much more.
BPCI Advanced will replace the Bundled Payments for Care Improvement (BPCI) initiative in October. CMS’s 44-page request for BPCI Advanced applications, released in January, was a breezy read compared with the densely detailed, 282-page final rule for the Comprehensive Care for Joint Replacement (CJR) bundled payment program.
|BPCI vs. BPCI Advanced
||Conveners and providers
||Conveners and providers
||Acute care hospitals, physician group practices, long-term care hospitals, inpatient rehabilitation hospitals, skilled nursing facilities, home health care agencies
||Acute care hospitals, physician group practices
|Included clinical episodes
||48 inpatient clinical episodes
||29 inpatient clinical episodes and three outpatient clinical episodes
||Model 2: Anchor inpatient stay + 30, 60, or 90 days of Part A and B
||Anchor acute stay or an outpatient procedure + 90 days of Part A and B (hospice services now included)
||2% or 3% depending on episode length
|Target price timing
||Retrospective target prices with adjustments for trend and wage index
||Prospective target prices with adjustments for case mix
||Composite quality score will adjust reconciliation payment amounts
“When CJR came out, you weren’t getting any sleep that night. When this came out, I read it three times and then watched TV for an hour,” quips Jonathan Pearce, principal at Singletrack Analytics, a New Jersey consulting firm specializing in financial technology and analytics.
The request for applications left many questions unanswered, but observers don’t expect that to dampen enthusiasm for BPCI Advanced. Interest in the program is strong among hospitals, medical groups, and what CMS calls conveners—groups that bear and apportion financial risk for providers. But until CMS releases more detail on pricing targets and other not-so-trivial nuances, providers who jump in may be taking a leap of faith.
What’s the target?
BPCI Advanced is a voluntary, five-year program that qualifies as an advanced alternative payment model under MACRA’s Quality Payment Program. Participants take on two-sided risk from Day 1—unlike CJR and the Oncology Care model, which allowed a couple of “practice” periods before providers were on the hook for losses.
BPCI Advanced includes 32 episodes spanning 90 days, including three outpatient-only episodes. The creation of the outpatient episodes means that physician groups will be able to “own” the risk for an episode in a Medicare bundled payment program for the first time.
“That’s a change we think is good, because so much of that episode is driven through physician groups and hospitals,” says Clay Richards, CEO of NaviHealth, a postacute care management company that assists hospitals with BPCI.
Payment reconciliation will hinge on quality performance, although the choice of measures “missed the mark,” said Amol Navathe, Ezekiel Emanuel, and colleagues at the Leonard Davis Institute of Health Economics in a Health Affairs blog post in February. The readmissions measure, they posited, parallels incentives in a bundled payment arrangement and does not change practice patterns, while another, advanced care planning, is irrelevant for some episodes.
As with its predecessor, BPCI Advanced target prices will be based on historical claims rather than the regional pricing method employed in the joint replacement program. That’s wise, says Pearce, whose consulting firm helps hospitals, ACOs, and other providers manage new payment models. “What that means is that you get paid for getting better, not being better, so it actually creates savings. If you use regional rates, somebody who gets paid less than the regional average doesn’t have to reduce costs at all.”
But target pricing is where we begin to enter the gray zone. As of early February, CMS had released no details on how target prices for episodes will be computed, other than to say that they will be prospectively determined on a preliminary basis, then retrospectively adjusted on the basis of cost trends and case mix.
“Everybody wants prospective targets, but a preliminary prospective target doesn’t really help you that much,” says Pearce. “If it’s different from the final prospective target by more than a couple of percentage points, which it usually is, that can be a huge swing in your dollars.”
Indeed, one of the chief complaints from participants in the original BPCI program was that performance targets were constantly changing—and worse, they didn’t know what their new targets were until after the start of the performance period. “Getting as much stability and predictability in that price will be important,” says Richards.
Predictability has particularly been a problem in the Oncology Care Model program. An editorial in the Journal of Oncology Practice last year pinned faulty target prices in the program on limitations in CMS’s risk-adjustment methodology. Though CMS has talked in the past about risk-adjusted targets in BPCI Advanced, the request for application is silent on this topic.
When the original BPCI program was launched in 2013, potential participants could ask CMS for claims data before applying—giving them a priceless look at where postdischarge costs might jump the shark. This time around, providers won’t get that information from CMS until after submitting an application—meaning they’ll have precious few weeks to study those data and choose which episodes, if any, to participate in. The wrong decisions may be more costly under BPCI Advanced than they were under the original program. In BPCI you could withdraw from an episode with 90 days’ notice, but in the new program, you’re stuck with the episodes you pick for a minimum of 15 months.
Hospitals currently in BPCI may have the advantage of using the data CMS shares with them, but there’s a caveat: CMS’s data-use agreements with BPCI providers say the information must be destroyed when the program ends because those data are personally identifiable. Pearce hasn’t yet heard any discussion about whether BPCI hospitals moving straight into BPCI Advanced will be able to keep their data.
“We’ve got three or four dozen hospitals in BPCI, and when we send them an email at the end of BPCI that says ‘you’ve got to delete everything we’ve ever given you, and we’ve got to blow away all your data on our servers and we can never go back to it,’ I think they’re going to say ‘Wait a second! We’ve participated in this program for five years. This shows what our results are. You mean we’ve got to get rid of this stuff?’”
Outpatient joint replacement missing
Among the 29 inpatient episodes, which are most ripe for savings? Maybe all of them, says Richards, whose NaviHealth participates as a convener for all 29 episodes in BPCI. Each of these episodes varies widely in postacute care costs, which Richards says range from 40% to 70% of total expenditures in an episode. “We’ve seen opportunities across the board.”
CMS included only three outpatient episodes—percutaneous coronary intervention (PCI), cardiac defibrillator surgery, and back and neck procedures other than spinal fusion. But total joint replacement—which was by far the most popular bundle in BPCI and, according to a Lewin Group analysis last October, was the one with the highest average savings compared with controls ($1,273 per episode)—was excluded. Both Richards and Pearce express surprise at the absence of a bundle for outpatient joint replacement, and Pearce thinks its omission could be problematic.
Why? Look to an unrelated decision by CMS last November, when the 2018 Outpatient Prospective Payment System final rule removed total knee replacement from the “inpatient only” list. With CMS no longer requiring that knee replacements be performed in the hospital, it is believed that uncomplicated cases will migrate to the outpatient side.
The problem, says Pearce, is that baseline costs in BPCI Advanced will be figured on claims from a time when all knee replacements—regardless of their complexity—were done on an inpatient basis. Pearce, whose company crunches data from the Comprehensive Care for Joint Replacement program for about 80 hospitals, used CJR data to estimate the effect of potential outpatient migration.
“I took out the bottom 20% of CJR episodes to simulate what would happen if 20% of your easiest patients went away. We recomputed the cost, and it was approximately $4,000 more per episode. Well, $4,000 more is going to kill you.”
That, he says, is why hospitals in the CJR’s voluntary metropolitan statistical areas are beginning to pull out of the program.
For all the uncertainties, there’s one clear lesson from BPCI that will serve providers who try BPCI Advanced: Go all in.
“When this began in 2014, you had to have a willingness to embrace care redesign, especially for postdischarge planning and follow-up. That requires change management and a real commitment from your executive leadership,” says Richards. “The benefit we’ve seen is that we’ve used bundled payments as the catalyst for care-redesign improvements for [all] payers, extending beyond BPCI.”