Oncology Rounds

The new 4-letter word: Attribution in the Oncology Care Model—and how CMMI is responding

by Deirdre Saulet and Soleil Shah

In conversations with Oncology Care Model (OCM) participants, we've heard consistent concerns about patient attribution, leading to it being dubbed the new four-letter word in cancer payment reform. Fortunately, the Center for Medicare and Medicaid Innovation (CMMI) has been responsive to stakeholders' feedback.

Check out our resources on CMMI's Oncology Care Model

Here's what you need to know about two new methodology changes for the OCM.

CMMI responds to concerns at annual stakeholder meeting

In early May of 2018, CMMI hosted its annual Oncology Care Model Stakeholder Meeting. During this public forum, participants shared concerns over various aspects of the OCM, including attribution and payment methodology. CMMI discussed strategies to address these issues, which include two new methodology changes:

  1. Attributing patients only to practices with at least one oncologist; and

  2. Refined target prices for breast cancer episodes.

Both changes will take effect for patients with episodes beginning after July 1, 2017, or Performance Period 3, and moving forward.

Here's our take on these changes—and what they may mean for OCM participants:  

OCM methodology change #1: Only attributing patients to practices with at least one oncology provider

Initially, CMMI attributed OCM patients to whichever practice conducted the plurality of a cancer patient's visits during a given performance period—regardless of whether the patient received cancer care at that specific organization, or if the practice offered any cancer care services at all.

This approach left a lot of room for misattribution. For instance, a cancer patient may see a non-oncology provider at Smith Hospital, but go to Jones Hospital for cancer care. This patient could still be attributed to Smith Hospital if he or she had more visits with his or her non-oncology provider than his or her oncology provider during the performance period. This method of attribution is bad for both Smith and Jones Hospitals if they're in the OCM: Smith Hospital is now held responsible for a patient's costs and outcomes, even though the hospital did not provide cancer care to that patient. Meanwhile, Jones Hospital—assuming that this patient was attributed to its practice—will likely have to pay CMMI back for the monthly enhanced oncology services (MEOS) payment the hospital billed for that patient.

To address this issue, CMMI will now attribute a patient to a practice only if that practice has at least one oncology provider. However, while this approach may narrow the pool of potential practices that can be considered to have the plurality of visits, it doesn't completely solve the attribution challenge. For instance, as long as Smith Hospital has at least one oncology provider, the patient could still be attributed to that hospital, even if he or she only received cardiology care there and went to Jones Hospital for cancer care. </p>

Ideally, only visits in which patients were treated by an oncology provider would count toward the "plurality of visits" requirement. But since providers self-report their specialty—and CMS does not attempt to verify this—provider classification is notoriously unreliable. Time will tell how CMMI will continue to improve attribution methodology in the OCM.

OCM methodology change #2: New target prices for breast cancer episodes

In the past, CMMI set target prices for breast cancer episodes in the OCM based on whether the episode included Part B chemotherapy drugs. Under this approach, episodes in which patients received any chemotherapy billed under Part B—whether or not they also received any drugs through Medicare Part D—were deemed "high-risk," while episodes in which patients received only Part D chemotherapy drugs were classified as "low-risk."

CMMI set lower target prices for low-risk breast cancer episodes than for high-risk episodes because the center assumed that low-risk episodes would primarily include cheaper, long-term oral endocrine therapies, such as anastrozole. However, more expensive Part D drugs that are not long-term oral endocrine therapies, such as palbociclib, are now more frequently used to treat breast cancer. Consequently, OCM participants found that the target prices for many of their supposedly "low-risk" breast cancer episodes were set too low to be feasible.

In response to this concern, CMMI changed its definitions for low-risk and high-risk breast cancer episodes. Now, any breast cancer episode in which a patient receives at least one of four specific Part D drugs—anastrozole, exemestane, letrozole, or tamoxifen—without any other chemotherapy is considered low-risk. Meanwhile, any breast cancer episode in which a patient receives at least one chemotherapy other than these four designated drugs is considered high-risk.

This change will likely help set fairer target prices as more expensive Part D breast cancer therapies enter the market. However, OCM participants still have at least two concerns about episode target prices that CMMI has yet to address:

  1. Merely changing episode definitions for breast cancer will not be enough to allay concerns over rising drug costs across all tumor sites. If target prices for an episode with rising drug costs stay fixed, cancer programs will find it increasingly difficult to spend less than the episode's target price; and

  2. Additionally, practices are still frustrated with adjustments for novel therapies, which they feel are not being accurately identified and incorporated into target price adjustments.
 

See first year performance results from the Oncology Care Model

Register for one of our 2018 National Meetings to hear more OCM updates.

Register Now

Learn more: Check out our resources on CMMI's Oncology Care Model

Get a comprehensive, constantly updated list of our resources on CMMI’s Oncology Care Model, which aims to improve care for chemotherapy patients while reducing costs.

Access the Resources