Characteristics of Hospitals Participating in the Transforming Episode Accountability Model

TL;DR

This study compares hospitals required to join a bundled payment program with those that are not, looking at various characteristics like spending and patient data.

Contribution

The study provides a detailed comparison of hospital characteristics between those mandated and not mandated to join the TEAM program.

Findings

Hospitals mandated to participate in TEAM differ in market and spending characteristics.

There are notable differences in patient demographics between participating and non-participating hospitals.

Abstract

This cross-sectional study compares the hospital-, market-, patient-, and spending-level characteristics of hospitals mandated vs not mandated to participate in the Transforming Episode Accountability Model (TEAM) bundled payment program.

Tables

Table 1.. Characteristics of Hospitals Participating and Not Participating in TEAMa

Characteristic	TEAM hospitalsb	Nonparticipant hospitals	P valuec
Hospital level
Total No. (%)	716 (100)	2046 (100)	NA
Beds, mean (SD) No.	263.0 (253.6)	230.0 (230.3)	.001
Urban location, No. (%)	684 (95.5)	1994 (97.5)	<.001
Ownership, No. (%)
For-profit	126 (17.6)	482 (23.6)	<.001
Not-for-profit	506 (70.7)	1336 (65.3)	.009
Public	84 (11.7)	228 (11.1)	.67
Geography, No. (%)
Northeast	211 (29.5)	217 (10.6)	<.001
Midwest	90 (12.6)	574 (28.1)	<.001
South	234 (32.7)	857 (41.9)	<.001
West	181 (25.3)	398 (19.5)	.001
Teaching hospital (major or minor), No. (%)	405 (56.6)	1030 (50.3)	.004
Member of a system, No. (%)	559 (78.1)	1586 (77.5)	.76
Safety-net hospital, No. (%)	99 (13.8)	177 (8.7)	<.001
DSH percentage points, mean (SD)	32.9 (18.7)	29.2 (16.9)	<.001
Hospital total margin, mean (SD)	0.01 (0.13)	0.02 (0.18)	.12
Hospital operating margin, mean (SD)	0.00 (0.15)	0.01 (0.22)	.07
County level
Census population in 2020, mean (SD) No.	694 185 (752 568)	1 071 448 (2 027 134)	<.001
Household income in 2021, mean (SD) of the median, $	76 023 (23 830)	67 201 (15 487)	<.001
Medicare Advantage penetration in 2022, mean (SD) %	45.6 (12.8)	46.5 (12.2)	.10
SNF total beds in 2022, mean (SD) No.	2754 (2753.8)	4399 (8291.6)	<.001
Rehabilitation hospitals in 2021, mean (SD) No.	1.01 (1.24)	2.50 (4.36)	<.001
Hospital market share, mean (SD) proportion	0.42 (0.38)	0.44 (0.40)	.10
Patient level
Dual Medicare-Medicaid enrollment, %	25.9 16.7)	23.3 (17.2)	<.001
Black race, mean (SD) %d	8.6 (12.3)	8.0 (12.1)	.26
Hispanic ethnicity, mean (SD) %d	6.4 (9.7)	5.8 (11.6)	.88
Annual discharges, mean (SD) No.	2162.0 (2408.47)	1894.3 (2033.73)	.004
Inpatient bundles
CABG, mean (SD)	13.4 (28.6)	13.0 (25.9)	.78
LEJR, mean (SD)	59.1 (144.54)	43.9 (61.92)	<.001
Major bowel procedure, mean (SD)	27.7 (38.6)	25.6 (33.7)	.17
Spinal fusion, mean (SD)	8.5 (16.3)	7.7 (13.4)	.17
SHFFT, mean (SD)	36.1 (38.1)	32.5 (35.7)	.02

Table 2.. Overall and Subcomponents of Adjusted 30-Day Episode Spending at Hospitals Participating and Not Participating in TEAM

Spending	Hospitals, $		P valuea
	TEAM participants	Nonparticipants
Overall 30-d episode
CABG	49 551	46 968	<.001
LEJR	28 315	25 914	<.001
Major bowel procedure	31 820	29 326	<.001
Spinal fusion	34 092	31 587	<.001
SHFFT	38 271	35 447	<.001
30-d PACb
CABG	5457	5486	.68
LEJR	9397	8813	<.001
Major bowel procedure	4368	4308	.23
Spinal fusion	6763	6416	.005
SHFFT	17 340	16 269	<.001
30-d SNF
CABG	2090	1718	<.001
LEJR	5708	5068	<.001
Major bowel procedure	2756	2368	<.001
Spinal fusion	2596	2075	<.001
SHFFT	11 951	10 157	<.001
30-d IRF
CABG	2574	3020	<.001
LEJR	2638	2996	<.001
Major bowel procedure	1164	1557	<.001
Spinal fusion	3834	3987	.19
SHFFT	5124	5869	<.001
30-d HHA
CABG	793	749	<.001
LEJR	1051	749	<.001
Major bowel procedure	449	384	<.001
Spinal fusion	334	354	.03
SHFFT	265	243	<.001

Full text

Introduction

The Centers for Medicare & Medicaid Services (CMS) recently announced the Transforming Episode Accountability Model (TEAM), a mandatory bundled payment program. Under TEAM, hospitals in selected US regions are accountable for spending on episodes of care, spanning admission to 30 days after discharge, for 5 surgical procedures.^1^ If hospitals meet spending targets, they receive a financial bonus from CMS; if not, they must pay CMS a penalty.

Prior bundled payment models were voluntary, allowing hospitals to start and end participation at will. This voluntary system was associated with selective enrollment by hospitals with high operating margins and low shares of marginalized patients.^2^ The selection bias, in turn, preceded payout of bonuses that exceeded spending reductions under the model, precipitating net financial losses for CMS.^3^

Moving from voluntary to mandatory participation addresses selection bias and increases participation from safety-net hospitals, but it has its own challenges. Safety-net hospitals mandated to participate in the Comprehensive Care for Joint Replacement (CJR) bundled payment model reduced spending on a magnitude similar to that for other hospitals yet were disproportionately penalized because they had lower (harder-to-meet) spending targets.^1^ Because TEAM sets spending targets similar to those of CJR, there is concern that safety-net hospitals might also disproportionately perform poorly under TEAM. Therefore, we sought to understand the characteristics of hospitals mandated to participate in TEAM as CMS looks to address selection bias and more equitably distribute financial incentives under bundled payment models.

Methods

We compared characteristics of hospitals selected vs not selected for TEAM using χ^2^ and t tests. We also compared spending between these groups after adjusting for patient- and hospital-level characteristics (eMethods in Supplement 1). The Human Research Protection Office at Washington University approved this cross-sectional study and waived informed consent because it was not human participant research. We followed the STROBE reporting guideline.

We estimated episode spending for TEAM’s 5 inpatient bundles: coronary artery bypass graft (CABG), lower extremity joint replacement (LEJR), major bowel procedure, spinal fusion, and surgical hip and femur fracture treatment (SHFFT). Analysis was conducted between November 2024 and March 2025 using SAS software (SAS Institute).

Results

We identified 716 hospitals participating in TEAM (Table 1). Compared with nonparticipants, participants were larger (263 vs 230 beds) and more often not-for-profit (506 [70.7%] vs 1336 [65.3%]), teaching (405 [56.6%] vs 1030 [50.3%]), and safety-net (99 [13.8%] vs 177 [8.7%]) hospitals. Participants were in wealthier counties (mean [SD] household income, $76 023 [$23 830] vs $67 201 [$15 487]) and served a higher share of patients with dual Medicare-Medicaid enrollment (25.9% vs 23.3%).

Participants and nonparticipants had similar episode volume for most bundles. Spending ranged from $28 315 for LEJR to$49 551 for CABG. For all bundles, 30-day episode spending was higher among participants than nonparticipants (Table 2). Among participants, postacute care (PAC) spending varied from $5457 for CABG to$17 340 for SHFFT, accounting for 11.0% and 45.3%, respectively, of 30-day spending.

Discussion

Hospitals mandated to participate in TEAM were larger, served more marginalized patients, and were often part of the safety net. This finding supports concerns that a large share of hospitals could bear disproportionate penalties under TEAM, as under CJR.^1,4^

Generally, TEAM participants had higher spending than nonparticipants, suggesting potential to reduce spending. However, this potential might vary across bundles. Participants in bundled payment generally lower overall episode spending by reducing PAC spending.^5,6^ Greater decreases might thus be expected in procedures for which PAC represents a greater portion of episode spending, such as LEJR and SHFFT bundles.^1,6^

Study limitations include the exclusion of 25 hospital participants that could not be linked to characteristics. We estimated spending for 5 inpatient bundles because they encompass most episodes under TEAM; TEAM also includes 2 outpatient bundles. Results of this study offer insight into the structural and spending characteristics of TEAM participants, including implications for equity and financial success under the model.