In October 2018, the Centers for Medicare & Medicaid Services (CMS) launched the Bundled Payments for Care Improvement Advanced (BPCI Advanced) program to continue expanding bundled payments nationwide. Participating hospitals and physician group practices are eligible for financial incentives if they meet quality and cost benchmarks for 90-day episodes beginning with either a hospital admission or an outpatient procedure. In a previous post and article, we described the 832 hospitals that began participating in October 2018.

Since then, CMS allowed a one-time opt-out in March 2019, which reduced the size of the hospital cohort to 715 hospitals. In January 2020, CMS enrolled a second cohort of participants and expanded the number of potential episodes from 32 (29 inpatient and 3 outpatient) to 35 (31 inpatient and 4 outpatient). Currently, 1,010 hospitals participate in BPCI Advanced.

In this article, we describe the organizational characteristics of hospitals participating in BPCI Advanced and the episodes they selected. We categorized participants into 670 “Wave 1” hospitals that entered in October 2018 and 340 “Wave 2” hospitals that entered in January 2020. We also identified 162 “Exiter” hospitals that left the program between October 2018 and January 2020.

Geographic Distribution

As shown in Figure 1, Wave 1 hospitals are spread across 193 regional health care markets (63% of total U.S. markets), while Wave 2 hospitals span 123 regional health care markets (40% of total U.S. markets). Collectively, Wave 1 and 2 hospitals represent 75% of all U.S. health care markets.

Exhibit 1: BPCI Advanced Hospital Participation by Market

Hospital Characteristics

Compared to Wave 2 hospitals, Wave 1 hospitals tend to be larger, teaching institutions located in urban areas and with higher hospital margins. Compared to current participants, exiting hospitals were small and more likely to be for-profit and non-teaching, with the lowest hospital margins and revenue paid on a shared risk basis.

Exhibit 2: BPCI Advanced Hospital Characteristics

Episode Selection

Overall, participating hospitals enrolled more frequently in inpatient than in outpatient or multi-setting episodes. Inpatient episodes were more frequently medical condition-based rather than procedure-based.

Selection of inpatient condition-based episodes varied by hospital group (Exhibit 3). Sepsis was the most commonly selected episode for both Wave 1 and 2 hospitals (62% and 50%, respectively), followed by COPD, bronchitis, asthma (39% and 28%, respectively) and cardiac arrhythmia (39% and 21%, respectively). On average, Wave 1 hospitals participated in more episodes than Wave 2 (5.7 versus 3.2 episodes). However, Wave 1 hospitals currently participate in fewer episodes than they did when they entered the program in October 2018 (5.7 versus 7.8 episodes). The most commonly selected episode among Exiters was congestive heart failure (47%). Other than for lower extremity joint replacement, the types of episodes selected did not vary meaningfully between Exiters and Wave 1 or 2 participants.

Exhibit 3: Inpatient Condition-Based Episode Selection

Similarly, hospital groups differed in selection of procedure-based episodes (Exhibit 4). Of note, few hospitals selected the new episodes introduced in January 2020, such as endovascular valve replacement, bariatric surgery, and inflammatory bowel disease.

Exhibit 4: Inpatient Procedure-Based Episode Selection

Hospitals selected outpatient episodes less frequently than inpatient ones. For instance, percutaneous coronary intervention was selected by 13% of Wave 1 hospitals, 9% of Wave 2 hospitals, and 15% of Exiting hospitals. Cardiac defibrillator was selected by <4% of hospitals in each group, and back & neck except spinal fusion by <6% of hospitals in each group. Major joint replacement of the lower extremity was introduced as a multi-setting episode in January 2020 but was not frequently selected (9% for Wave 1 and 8% for Wave 2).

Final Thoughts

Our findings highlight how entry into and exit from BPCI Advanced have changed both the types of hospitals participating and the types of episodes selected under bundled payments. Although program results will not be available for several years, our findings suggest that:

• Certain hospital attributes (e.g., experience with financial risk) may be pivotal for success in bundled payments. Exiting hospitals tended to be smaller, for-profit institutions with less revenue on a shared risk basis.
• Outcomes may be easier to manage for certain episodes than others. Exiters were more likely to select congestive heart failure episodes than current participants.

Authors’ Note

This study was funded by The Commonwealth Fund.