The Key Factors Driving Low Participation in the Enhancing Oncology Model

 J Clin Pathways. 2024;10(1):13-17. doi:10.25270/jcp.2024.01.01

The Enhancing Oncology Model (EOM) was designed to expand oncology’s foot­print in value-based care as a successor to the Oncology Care Model (OCM).¹ A week prior to the launch of the EOM on July 1, 2023, the CMS Innovation Center (CMMI) released the first participant list, consisting of 67 practices. However, only 12 days into the new model, CMMI released an updated list with just 44 oncology practices.² In addition to this steep attrition, the 44 EOM participants are notably fewer than the 190 practices that originally joined the OCM. What happened to cause this sharp decline?

I worked with 18 academic oncology practices that participated in the Association of American Medical Colleges (AAMC) OCM Collaborative. When the EOM appli­cation opened, 14 of these practices applied and were accepted into the program, but ultimately all chose to drop out of the model prior to its July 1 start date. The three central factors that all the practices cited as driving their decision not to participate in the EOM, despite having built enhanced treatment pathways and value-based care infrastructure through the course of the OCM, were:

1. immediate downside risk in the current climate of narrow and uncertain provider margins;
2. inadequate risk adjustment that does not sufficiently account for drug costs; and
3. onerous reporting burdens for data that may not be used.

In addition, some practices discussed broader strategic objectives—including par­ticipation in other risk-bearing value-based care programs with Medicare and private payers—as being more essential to prioritize given their limited operational and financial resources.

The EOM requires all participating oncology practices to take immediate downside risk for seven cancers: breast can­cer, chronic leukemia, small intestine/colorectal cancer, lung cancer, lymphoma, multiple myeloma, and prostate cancer. In contrast, the OCM originally allowed practices up to 3.5 years of upside-only risk and ultimately extended this option for the entire model in response to the pandemic. Practices are ac­countable for their patients’ total cost of care, including chemo­therapy and other treatment regimens, for 6-month episodes.

This design creates financial incentives to prescribe lower-cost drugs. However, oncologists select treatment regimens based on the patient’s specific characteristics, including cancer stage and molecular mutations. In many cases, the most clinically ap­propriate treatments require very high-cost drugs, resulting in good patient care but financial losses under the EOM.

For the academic OCM practices I worked with, the potential losses represented by this model were staggering. Under the OCM, the academic practices were more than $15 000 over target for the average lung and multiple myeloma episode (Figure 1). Since the practices were under upside-only risk, these “losses” were hypothetical because they were not required to pay CMMI back for spending that was over target. Had the practices taken on downside risk, CMMI would have reduced the losses by 4% under the terms of the model. However, even when all OCM episodes were consid­ered, the potential losses on lung and multiple myeloma out­weighed potential savings on all other cancers by a cumulative $85 million across the first 4.5 years of the model (Figure 2).

While the EOM made some methodological changes to the risk adjustment, many of the core model elements remain the same, especially the inclusion of the drugs that drive the cost of oncology care. In addition, the EOM eliminated many of the cancer types that achieved savings under the OCM, including low-risk breast and prostate cancer. Several academic practices conducted internal modeling that showed potential losses of several million dollars per year—far more financial risk than they were prepared to take, especially in the current context of escalating cost for supplies and labor.

In addition to the size of the financial risk, many of the aca­demic practices believed that the EOM risk adjustment was inadequate to account for appropriate drug regimens. This was also a concern under the OCM, which CMMI partially addressed in the EOM by applying sophisticated claims-based risk adjustment for individual cancer types. This risk adjust­ment accounts for the rising cost of drug treatments in aggre­gate, as well as providing a 2-year novel therapy adjustment for practices that more quickly implement new drugs than other oncology practices nationwide.

However, there is a crucial blind spot for this claims-based adjustment: The patient’s cancer stage, clinical status, and mo­lecular mutations cannot be accounted for without including clinical data. Since these clinical factors drive the decision-making for a given patient’s treatment regimen, failure to ac­count for this means that the EOM does not adequately adjust for drug costs within each cancer type. Instead, the risk adjust­ment simply assumes that each practice treats a normal distri­bution of patients, which is a fallacy on two counts. First, even high-volume practices do not see enough patients to achieve a normal distribution curve. Second, some practices dispropor­tionately draw patients with specialized treatment needs due to a patient’s genetic markers, cancer stage, cancer recurrence, or treatment history.

Without adequate risk adjustment, performance under the EOM will ultimately be driven by which drug regimen is prescribed, as it was under the OCM, because chemotherapy accounts for such a large portion of total oncology spending. Figure 3 and Figure 4 show that chemotherapy represented 70% of total multiple myeloma spending on OCM episodes (averaging $61 175 per episode) and 67% of total lung cancer spending ($45 684) by the second half of 2020. No amount of cost savings from reducing avoidable admissions and emer­gency department (ED) visits can achieve savings relative to these drug costs. Yet, even in their most recent evaluation of the OCM, CMMI staunchly supports the belief that there is opportunity for savings in the EOM by switching to lower-cost drugs for a given cancer.³ However, in many cases, the most clinically appropriate drugs are very high cost due to patient characteristics such as cancer stage and molecular mu­tations. Clinicians put the patient first by choosing the right drug, not the cheapest one, in cases where the cheapest one would be less effective.

CMMI can resolve this concern by incorporating clinical data into the risk adjustment methodology to account for significant clinical variables.⁴ CMMI is currently collecting extensive clinical data from the EOM practices about each pa­tient, as they did in the OCM. CMMI first introduced clinical risk adjustment near the end of the OCM model for meta­static status for breast, lung, and small intestine/colorectal cancers. Under the EOM, CMMI is applying a similar risk adjustment for “ever metastatic” cancers, which is a more in­clusive variable, as well as adjusting for HER2 breast cancers. While the use of this clinical data is promising, CMMI must go further to ensure adequate risk adjustment for all cancers in the model. For example, the EOM currently does not apply any clinical risk adjustment for multiple myeloma, nor does it incorporate other staging data in the risk adjustment for any cancer type.

Many practices are especially frustrated by the EOM’s insuffi­cient use of clinical risk adjustment, given that CMMI requires practices to report on a much more extensive list of clinical data elements for each episode related to the patient’s cancer stage, clinical status, and molecular mutations. To collect this data, oncology practices must either build and populate machine-readable fields into their electronic medical record, or they must manually abstract this data from their clinical documen­tation. This administrative burden is doubled by the fact that practices must also report the same elements to cancer regis­tries, but CMMI will not use cancer registry data because it does not align with their timeline.

The EOM reporting requirements are similar to those un­der the OCM, which practices struggled to achieve despite spending years refining the data-collection process. To cover the labor and technological costs of this reporting requirement, many OCM practices used funds from the monthly enhanced services payment, which were set at $160 under the OCM but dropped to $70-$100 under the EOM. While the academic practices agree that clinical data is essential for adequate risk ad­justment, they believe CMMI should only require practices to report on data elements that the agency uses for either financial or quality measurement.

CMMI has several policy options available to address the meth­odological concerns that have depressed EOM participation. Firstly, CMMI should reconsider the level of financial risk that oncology practices are required to assume under the EOM. The agency has a number of options available to mitigate financial risk, including:

• Reducing the cut of any savings retained by CMMI (referred to as the Centers for Medicare & Medicaid Services discount) and introducing risk corridors for losses over certain amounts
• Only requiring practices to take on financial risk for the nondrug portion of the episode, making practices financially responsible for appropriate inpatient and ED utilization but not for chemotherapy
• Reducing downside risk for practices with high-quality performance (as occurs under CMMI’s Bundled Payments for Care Improvement Advanced Model)

Secondly, CMMI should refine the EOM target price methodology to incorporate the clinical data they are already requiring practices to report. This would help mitigate the concerns about inadequate risk adjustment within each cancer type and, more appropriately, adjust for the cost of drug treat­ment regimens.

Thirdly, CMMI should reduce the reporting burden for clinical data by only requiring practices to report the data that is used in the EOM financial and quality methodologies. CMMI should also explore whether cancer registries could provide the data elements they require, which would remove the need for practices to report duplicative data to two organizations.

Collectively, these refinements would strengthen the EOM for all oncology practices—whether academic or community based—by more accurately measuring performance and re­warding oncology practices for delivering high-quality, appro­priate patient care.

Authors: Theresa Dreyer, MPH

Affiliation: Health Care Transformation Task Force, Washington, DC

Address correspondence to: 

Theresa Dreyer, MPH

5636 Connecticut Ave NW

PO Box 6258

Washington, DC 20015

Email: theresa.dreyer@hcttf.org

Disclosures: The author reported no financial or other conflicts of interest.