[Reposted: Amol S. Navathe, Justin E. Bekelman and Joshua M. Liao.  Advancing Bundled Payments Through Radiation Oncology Episodes, Health Affairs Blog, August 1, 2019. https://healthaffairs.org/do/10.1377/hblog20190729.824220/full/: Copyright ©2019 Health Affairs by Project HOPE – The People-to-People Health Foundation, Inc.]

On July 10, 2019, the Centers for Medicare and Medicaid Services (CMS) announced a proposal to test bundled payments for radiation oncology services for 17 types of cancer. Although it shares features with existing bundled payment programs, the Radiation Oncology (RO) Model is positioned to launch CMS payment policy forward with novel features aimed at curbing known inefficiencies in spending on radiation therapy. 

Features Shared With Preceding CMS Bundled Payment Programs

As proposed, the RO Model shares or extends features used in preceding bundled payment programs. Like those in the ongoing Oncology Care Model (OCM), episodes in the RO Model would be triggered by initiation of therapeutic service—in this case, radiation therapy. Like the Comprehensive Care for Joint Replacement (CJR) Model, the RO Model requires participation based on location in randomly selected geographic markets. The RO Model also ensures that CMS achieves savings by setting a discount factor (that is, the percentage by which CMS reduces payment for an episode, relative to traditional fee-for-service amounts) and sets payment based on national base rates adjusted for each participant’s geographic location, patient panel, and historical treatment patterns. 

Like the OCM and CJR Model, the RO Model works by adjusting financial savings or losses based on performance on a small number of quality metrics. Finally, the RO Model follows the precedent from other recently implemented bundled payment programs by qualifying as an advanced alternative payment model (APM) under Medicare’s Quality Payment Program. 

Despite these similarities, the RO Model is also poised to drive value by pulling three “policy levers” that have either been partially and wholly unexplored in prior programs. 

Lever 1: Prospective Payment In An APM

CMS intends to pay RO Model participants on a prospective, rather than retrospective, basis using separate payments for professional and technical components of care. Building upon several smaller demonstrations trialed over the past decade, this design represents the agency’s largest scale foray into prospective payment in a formal APM. 

This move toward prospective payment increases the stakes and complexity of the arrangement for both Medicare (that is, moving away from fee-for-service systems by dispersing payments upfront) and participants (that is, accepting fixed sums pre-adjusted for organizational, regional, and national factors). 

In its most mature form, prospective payment may reduce some aspects of administrative complexity, for example, eliminating the need for providers to submit individual billing codes required for fee-for-service. However, it could create other administrative needs, such as adjusting payments for beneficiaries that switch radiation oncology providers mid-episode. There may also be other effects: Prospective payment may increase psychological alignment to reduce low-value services by leveraging the behavioral principle of “loss aversion,” which describes individuals’ tendency to prefer avoiding losses (in this case, payments already received from Medicare) over achieving equivalent gains. 

Lever 2: Site Neutrality

Site neutral payments are a key feature of the RO Model. In the press release and proposed rule, CMS explicitly identified site-of-service payment differentials between freestanding centers and hospital outpatient departments as one major driver for implementing the model for radiation oncology. While freestanding centers are historically paid lower rates, they provide a higher volume of radiation services and use more expensive modalities that more than offset those lower rates. In particular, based on an analysis of Medicare fee-for-service claims, CMS reports that radiation therapy episodes triggered at freestanding centers were paid over 10 percent more by Medicare compared to episodes triggered in hospital outpatient departments. 

As a result, CMS proposes addressing site neutrality using an episode-pricing model that reimburses participants based on hospital-outpatient department rates, irrespective of site of care. The episode payment for each type of cancer is based on a national base rate (which represents the historical average episode spending) and a trend factor (which reflects secular changes in payment outside of the payment model; that is, payment to non-participants in both hospital outpatient and free-standing center settings). Payments are also adjusted for each participant’s specific historical experience, geographic region, case-mix, and the above-mentioned episode discount factor. 

Taken together, these steps are designed to better align provider and beneficiary interests by removing fee-for-service incentives to provide greater intensity or complexity of radiation services and select more expensive sites of care. 

Lever 3: Jumping Straight To Mandatory Participation In An APM

While the RO Model is not the first mandatory Medicare APM—that distinction goes to the CJR Model, which CMS implemented in 2016—the proposed model represents Medicare’s first attempt to require APM participation without a preceding voluntary program. 

The CJR Model was designed based on the format and positive early results from its predecessor program, the Bundled Payments for Care Improvement initiative. The Shared Savings Program now requires all participating organizations to assume some degree of downside financial risk but only after more than five years of permitting participants to engage in financial upside-only arrangements. Going straight to mandatory participation represents a new approach for implementing APMs.  

As proposed, the model would also be mandated more broadly than preceding bundled payment programs. In the CJR Model, only hospitals in metropolitan statistical areas with high historical joint replacement spending were eligible for random selection into the program. In the RO Model, radiation therapy providers located in all core-based statistical areas (CBSAs)—including both metropolitan and less populated, but non-rural “micropolitan” areas—are eligible for the program. To reach a goal of sampling 40 percent of all relevant episodes nationwide, CMS proposes to stratify radiation therapy providers across all CBSAs by historical episode volume before randomly selecting participants within each stratum for participant and comparison groups. 


The recently proposed RO Model emphasizes several new policy levers that can drive health care value but have been partially or completely unexplored in bundled payment programs to date: prospective, site-neutral episode payments implemented nationwide on a mandatory basis. In turn, pending confirmation in a final rule, these features are important to consider as potential mechanisms for advancing payment reform across specialty models. 

Authors’ Note

Dr. Navathe reported receiving grants from Hawaii Medical Service Association, Anthem Public Policy Institute, Cigna, Healthcare Research and Education Trust, and Oscar Health; personal fees from Navvis Healthcare, National University Health System of Singapore, and Agathos, Inc.; personal fees and equity from NavaHealth; speaking fees from the Cleveland Clinic; serving as a board member of Integrated Services Inc. without compensation; equity from Embedded Healthcare; and an honorarium from Elsevier Press, none of which are related to this blog post. Dr. Bekelman reported receiving personal fees from CVS Health unrelated to this blog post and personal fees in 2017 and 2018 from Actuarial Research Corporation as a consultant on radiation oncology to the Centers for Medicare and Medicine Services Center for Medicare and Medicaid Innovation.