Bundled payments: Navigating the next era of value-based care

By: Avinash Prasad, Harbinder Raina and Talha Bhatti

Bundled payments are no longer a side experiment in Medicare reimbursement. They are becoming a central mechanism for shifting providers from volume-based payment to episode-level accountability. Under models such as TEAM, hospitals are responsible for the cost and quality of defined surgical episodes, with financial performance measured against CMS target prices. The result is a more demanding operating environment where health systems must manage clinical variation, post-acute utilization, readmission risk and patient outcomes with far greater precision.

This shift matters because bundled payments sit in the middle of the value-based care risk continuum. They are more financially consequential than pay-for-performance programs, but more operationally bounded than full capitation. That makes them a practical but unforgiving test of whether a provider organization can translate data, care coordination and clinical governance into measurable performance.

For many providers, the question is no longer whether to participate. Mandatory models are expanding, and the organizations that build stronger episode management capabilities now will be better positioned as CMS continues to move more reimbursement into accountable payment models. TEAM, CJR-X, ASM, ACCESS, LEAD and WISeR signal that episode-based accountability, chronic care management, specialist alignment and technology-supported oversight are becoming more prominent parts of the federal payment landscape.

Winning in this environment requires more than a dashboard or a retrospective claims review. Providers need a repeatable operating model that identifies the right patients early, stratifies risk accurately, guides discharge planning, directs patients to high-performing post-acute partners, and gives leaders near real-time visibility into episode economics. AI and advanced analytics can help, but only when they are embedded into the clinical and operational workflows that determine cost and quality at the point of care.

ZS helps health systems build this bridge between analytics and action. Our approach combines episode strategy, data infrastructure, AI-enabled risk and referral tools, care management workflow design and performance management capabilities that support both near-term execution and longer-term value-based care readiness.

1. Bundled payments change the economics of care delivery

Traditional fee-for-service reimbursement pays providers for each separate service delivered during a patient’s care journey. A physician visit, imaging study, implant, lab test, hospital stay and skilled nursing facility day are each reimbursed as individual transactions. This model rewards volume and does not require the provider to manage total episode cost across the full care journey.

Bundled payments change that logic. A single prospective or retrospective payment covers care from an anchor event through a defined post-discharge window, often 30 to 90 days. If the provider manages the episode below the target price while meeting quality requirements, it can retain a share of the savings. If the episode exceeds the target price, the provider may owe a repayment. In practical terms, the health system takes on both the upside and the downside associated with variation in care delivery, post-acute referrals, readmissions, and resource use.

This creates a different management problem for health systems. Success is not determined only by the procedure itself; it depends on whether the organization can standardize the care pathway, reduce avoidable variation, manage discharge disposition, coordinate post-acute care, and monitor patient risk after discharge. Taking hip replacement as an example, the opportunity comes from areas like eliminating duplicative imaging, optimizing implants, shortening skilled nursing facility stays when clinically appropriate, and managing the overall episode below the payer’s bundled amount.

2. Mandatory participation is raising the urgency

CMS’s recent model activity points to a clear direction of travel. TEAM requires selected hospitals to participate in 30-day surgical episodes across five procedure categories. CJR-X would expand mandatory joint replacement accountability across a much broader hospital footprint, with a 90-day episode window. Other models, including ASM, ACCESS, LEAD and WISeR, further reinforce CMS’s focus on specialty care management, chronic condition support, accountable care and utilization oversight.

TEAM is especially important because it makes episode accountability operationally unavoidable for participating hospitals. The model covers lower extremity joint replacement, surgical treatment for hip femur fracture, spinal fusion, coronary artery bypass graft and major bowel procedures. Hospitals in selected geographic markets are required to participate, and the model runs through December 31, 2030.

The TEAM methodology uses historical spending, regional adjustments and a CMS efficiency discount to establish target prices. Hospitals that outperform their target price can receive reconciliation payments, while those that overspend can face repayment obligations. Quality scores affect reconciliation outcomes, which means financial performance and clinical quality are directly linked.

CJR-X raises the stakes further. The proposed model serves as an expansion of mandatory joint replacement accountability, with a 90-day episode window and broader post-acute cost exposure. That longer episode window increases the importance of discharge planning, preferred post-acute networks, home health utilization, skilled nursing facility management and patient monitoring after discharge.

The implication for providers is straightforward: waiting for reconciliation data is too late. Organizations need to understand episode performance while there is still time to intervene.

3. Six operating challenges determine episode performance

First, many providers choose bundles without enough analytic rigor. Historical cost data, risk-adjusted benchmarks, case mix, surgeon-level variation and competitor context are needed to understand where the organization can reasonably perform below target prices. Without that analysis, providers may enter episodes where they have limited ability to manage cost or quality.

Second, patient identification is often too manual. If eligible patients are identified late, care teams lose the most valuable window for preadmission planning, risk stratification and proactive intervention. Automated identification using ADT and EHR feeds is becoming a baseline capability for effective episode management.

Third, clinical variation drives avoidable cost. Differences in implant selection, imaging frequency, length of stay, discharge planning and post-acute referral patterns can materially change episode economics. Standardized care pathways do not remove clinical judgment, but they do create a consistent baseline from which exceptions can be managed more deliberately.

Fourth, resources are often allocated too evenly. High-risk patients may not receive enough care management support, while lower-risk patients receive interventions they may not need. Risk stratification helps care teams decide where to focus outreach, follow-up, medication checks and escalation resources.

Fifth, post-acute care remains one of the largest controllable cost levers. Referrals to out-of-network or high-cost post-acute providers can materially affect episode profitability. Preferred post-acute networks and decision support at discharge are therefore critical to performance.

Finally, limited performance visibility prevents timely course correction. If leaders only understand performance after CMS reconciliation, they cannot address emerging cost or quality issues while the episode is still active. Continuous performance tracking is required to manage bundled payments as an operating discipline rather than a retrospective finance exercise.

4. AI can improve episode management when it is tied to workflow

AI is most valuable in bundled payments when it improves a concrete operational decision, and can be deployed across six capability domains: bundle selection, patient identification, risk scoring, referral optimization, care management and analytics. These domains provide a practical way to think about where AI can improve performance.

In bundle selection, predictive models can analyze historical episode cost, case mix, surgeon variation and regional benchmarks to estimate expected financial performance. This helps providers decide which episodes to prioritize and which operational improvements are required before taking on more risk.

In patient identification, rules-based engines and natural language processing can monitor ADT feeds, EHR data and order streams to identify eligible patients earlier. Earlier identification gives care teams more time to coordinate preadmission planning, flag high-risk patients, and prepare discharge pathways.

In risk stratification, AI can combine clinical, social and behavioral variables to identify patients at higher risk of readmission or avoidable utilization. These insights can drive targeted outreach, follow-up scheduling, medication adherence support and escalation to clinicians when needed.

In referral optimization, AI can help discharge planners compare post-acute providers based on cost, quality, network status and geography. This moves referrals away from purely relationship-based patterns and toward evidence-based site-of-care decisions.

In care management, intelligent worklists can prioritize outreach based on missed follow-ups, medication adherence signals, predicted deterioration, and other risk indicators. This helps care coordinators focus their time where it is most likely to reduce avoidable utilization.

In performance analytics, dashboards can combine claims and EHR data to identify cost outliers, physician variation, skilled nursing facility utilization patterns, network leakage, and projected financial exposure. Generative AI can help summarize performance drivers in plain language, but it should support decision-making rather than replace clinical or operational judgment.

5. The ZS approach: connect analytics to execution

Our perspective is that providers do not need another disconnected analytic tool. They need an integrated operating model that turns episode data into timely action. Providers can follow a four-layer framework that connects the data foundation to performance management.

The first layer is the data and analytics foundation. Providers need to ingest and harmonize EHR, ADT, CMS claims and post-acute care data into a usable episode analytics environment. This enables historical cost benchmarking, real-time patient identification and the model inputs required for risk scoring and referral optimization.

The second layer is intelligent episode operations. This includes automated patient identification, AI-enabled risk stratification, preferred post-acute provider lists and patient-reported outcome data capture where required. These capabilities help move bundled payment management upstream, before cost and quality issues are locked in.

The third layer is care management and coordination. Standardized protocols, role-based workflows and patient-facing communication help translate analytics into actions taken by care coordinators, physicians, post-acute partners and patients.

The fourth layer is performance management and strategy. Leaders need episode-level financial dashboards, quality composite tracking, network leakage analysis, physician variation reporting and scenario modeling for future model participation. These tools help the organization decide when to expand, where to intervene, and how to prepare for models such as CJR-X.

6. A practical readiness agenda

For providers participating in TEAM or preparing for future bundled payment expansion, the next 12 months should focus on execution.

In the foundation phase (Q1), providers should assess data infrastructure, analyze historical episode costs, review target prices, establish baseline quality performance, and define their convener strategy. In the operations build phase (Q2), they should integrate ADT and EHR feeds, automate patient identification, build risk stratification models, define care protocols, and establish preferred post-acute networks.

Once those capabilities are in place, organizations can shift toward performance lift (Q3). This includes launching episode dashboards, reviewing physician variation, analyzing network leakage, simulating reconciliation performance and assessing the likely impact of expanded joint replacement accountability. Over time, the same infrastructure can support broader specialist alignment, ACO integration, generative AI workflow support and commercial bundle strategy (Q4 and beyond).

Providers should also establish a focused set of success metrics. These could include risk-adjusted cost per episode, in-network referral rate, 30-day readmission rate and home health utilization ratio. These metrics give leaders a practical view of financial performance, referral discipline, avoidable utilization and post-acute care mix.

Bundled payments are becoming a more prominent and more demanding part of the value-based care landscape. TEAM is already active for selected hospitals, CJR-X points toward broader joint replacement accountability, and CMS’s broader model activity suggests that episode and specialty accountability will continue to expand.

For providers, success will depend on building durable operating capability before financial exposure increases further. That means stronger data infrastructure, earlier patient identification, more precise risk stratification, better post-acute referral management, standardized care pathways and performance visibility that supports in-flight intervention.

The organizations that invest now will be better positioned to protect margin, improve quality, and compete in a reimbursement environment where value-based accountability is no longer optional. ZS can help providers accelerate that journey by connecting episode strategy, AI-enabled analytics, care management workflow and performance management into one practical operating model.

Read more insights from ZS.

This article reflects the authors’ personal views and does not necessarily represent any official position of ZS.