Operational Innovations, AI to Play Vital Roles in EOM Data Collection

Barry Russo, MBA

The arrival of the Enhancing Oncology Model (EOM) has renewed the uncertainties surrounding value-based care approaches for practices opting into the next wave of Center for Medicare & Medicaid Services (CMS) performance structure. Touted as building on the lessons learned under the Oncology Care Model (OCM), implemented in 2016 and expired in 2022, the goals of the EOM are structured around keeping patients at the center
of the structure with increased engagement for all parties involved with their care and a renewed focus on improving health equity and savings.

However, lessons learned from practices that participated in the OCM and improvements made in data collection and analysis stand to usher in a new wealth of informative approaches for optimizing cancer care delivery.

“The biggest concerns that practices have are understanding and predicting how they’ll perform and the likelihood that they will achieve performance payments,” Amy Valley, PharmD, vice president for clinical strategy and technology solutions at Cardinal Health said in an interview with OncologyLive. Valley added that the most significant barrier for practitioners having confidence going into the EOM is understanding the analytics of their practice. “Practices need to be comfortable accepting downside risk, right out of the gate. If you don’t understand your analytics and how it relates to your patient population, it’s hard to take on that risk.”

One of the biggest takeaways from conversations surrounding the EOM is not just the volume of data ahead of participation, but also the amount that will emerge once the participa- tion period begins. “When [Cardinal Health] was doing market research about the unmet needs and persistent issues of the OCM, we identified that practices weren’t getting their data files until well past an episode or performance period had closed,” Valley said. “It was cumbersome to work with the data, take away insights, and then make process improvements. As a result, in the OCM, we saw delayed but improved performance.”

Once the data arrive, Valley noted that it may not be intuitive for practices to take actionable steps toward improvement. “Even though CMS is providing more regular data files, [it needs] to be in a format that can easily be understood, know what your drivers are, know when something [is off] or some new problem bubbles up, and be able to act quickly. The need for prompt data and actionable insights is going to continue to be a very important solution for practices to have to be successful in any sort of value-based reimbursement program.”

Under the EOM, CMS noted that partici- pants are incentivized to engage proactively through involvement in several areas including providing 24/7 access to care, evidence-based guideline approaches, collecting electronic patient-reported outcomes (ePROs), and patient navigation services. In terms of improvements for closing gaps in health disparities, social determinants of health will play a larger role in patient evaluation and risk mitigation.¹

For example, EOM participants are required to submit health equity plans outlining strategies for their patient population. During engagement with patients, clinicians are tasked with screening for health-related social needs.¹

Barry Russo, MBA, CEO at The Center for Cancer and Blood Disorders in Fort Worth, Texas, noted that for practices that participated in the OCM, the practice improvement structures that were required under that model should have laid the groundwork for participation the EOM. “With OCM, we wanted to be successful, we wanted to do the right thing and understand how to manage a population,” Russo said in an interview with OncologyLive. “As we move to the EOM, it’s even more important to understand infrastructure needs because you’re at full financial risk for the patients who are attributed to you.”

In tandem with infrastructure needs, the narrowed focus on disease states and the amount of data being collected in the EOM may make a difference in reporting. Included under the EOM are patients receiving systemic chemotherapy for solid tumors including breast, small intestine/colorectal, lung, and prostate cancers, and hematologic malignancies including chronic leukemia, lymphoma, and multiple myeloma.²

“The number of disease states is down to 7, as opposed to the OCM, which was pretty much anyone who had cancer,” Russo said. “You’re going to have [fewer] patients attributed, which makes the risk higher, but for those who participated in OCM, an infrastructure is already in place. What we don’t yet have completely nailed down are the new social determinants of health requirements. We need to create the mechanisms to not only track and report on them, but also understand the implica- tions that [these data] add to managing [our] patient populations.”

Collection and Analysis of New Data Points

Russo added that social determinants of health data and ePROs pose a new challenge not only in data collected but also in how those data are collected. “What systems are we going to use and how are we going to manage that? What do we do with the information coming back from the patient, and how are we acting and reacting to it?” he asked.

Quality improvement data, patient-level clinical data, staging data, and sociodemographic data are required to be reported once per performance period.² “By merging the clinical and social data, there are a lot of social deter- minants of health end points to risk-stratify patients,” Russo said.

Benefits of using ePRO tools in oncology settings have been reported as leading to streamlining processes, promoting patient- physician conversations, and improving engagement.³ The tools recommended for ePRO collection under the EOM include screen-based reporting devices (ie, patient portal on a smart- phone or computer), interactive voice response systems, SMS text, or collection via tablet in a waiting room.³ The EOM is using a grad- ual implementation approach for collection but anticipates that by year 5 of participation 75% of EOM-specified beneficiaries will have ePROs documented in subsequent visits.

The influx of data collected from ePROs and data input into electronic medical records will likely increase the burden on an already overtasked clinical workforce. Ensuring that platforms “talk” with one another and streamline data is the first step to finding ideal uses for artificial intelligence (AI)-driven tools to make sense of the data coming from the practice and CMS.

“Connectivity between these systems is necessary to reveal the insights that will lead to meaningful interventions for patients,” Valley said. “Practices must be smart about arming their workforce with tools that make the practice as efficient as possible, especially as practices struggle with workforce reten- tion and competition of precious resources. Utilizing AI in this area has great promise to deliver a meaningful return on investment for a practice.”

Cardinal Health’s Navista TS platform has offered integrated AI tools since its launch in 2021, using Jvion AI, which was acquired by Lightbeam Health Solutions, to help iden- tify patients at risk for hospitalizations or emergency department visits.⁴ Leveraging this collaboration, the platform can prioritize patients based on ePRO responses. Considering the new requirements and areas of data collection required by the EOM, the platform has been optimized to address the 2-sided risk model to call attention to social determinants of health and comorbidities aligned with high-cost targets, calculate total patient costs, and supply alternative ePRO collection touchpoints.

“Practices who participate in the EOM will need to figure out how to proactively moni- tor their entire patient populations,” Valley said adding that tools need to be able to generate actionable findings for patients at rising risk and, more specifically, those who are clas- sified under the 7 tumor types in EOM and need to be closely monitored. “Patients may use ePRO tools to rank their pain levels, symp- toms like nausea, or even indicate a call into the practice. Practices need a command center so that they can see their entire patient population, and engage with them based on their treatment plan, or their acuity. The data inputs from ePROs can help alert practices of patients at rising risk so that patients may be triaged appropriately. Applying AI is going to be crit- ical to these massive amounts of data that are being generated. You’ve got to have an engine to be able to sort through all that so you can prioritize and act on it.

“As these tools are validated, they will be more widely adopted,” Valley said. “And as they are more widely adopted, the cost of the innovative technologies will drop. This is good news for the community oncology practices looking to take advantage of [these technologies].”

Russo noted that his practice helped pilot the tool from Cardinal Health under the OCM and it plans to leverage the AI technology moving forward in the EOM. “I’m hoping there’s more AI that will assist us in recognizing issues or trends out of claims data. That would be helpful to us,” Russo said.

Lessons Learned from the OCM for the EOM

Russo noted that under the OCM, several key takeaways from The Center for Cancer and Blood Disorders’ participation shined through in 4 areas: education, care delivery, importance of data, and communication. “You cannot over-educate,” Russo said. “We now make it a point that value-based care and EOM-related items are on the agenda of every committee meeting we have inside the practice.” Regarding care delivery, Russo stressed the multidisciplinary nature of the referral process. “When you’re managing a population, you don’t just make a referral, you manage it, ensuring the patient gets [to the appointment], that they get there in a reasonable period of time, and, in some cases, this means you must develop a relationship with pulmonologists, cardiologists, endocrinolo- gists, etc,” Russo explained.

In addition to the notes on data surrounding ePROs, Russo highlighted that understanding claims data plays an important role in manag- ing a practice. “The third thing we learned with the OCM is that data, especially claims data, are important. You get a claim for every provider, hospital, any health care event a patient has where a bill was sent in, which gives you a much larger, longitudinal view of what’s happening with the patient, which helps you in the overall management of the patient and provides an understanding for how to process claims data.”

Finally, Russo noted that finding the right way to encourage patient communication was vital to success under the OCM. “We were more successful, the more often we communicated with our patients. You can’t do that by just picking up the phone. No one wants to talk on the phone anymore. They may just want to text message. We must have an electronic communication method with the patient. For patients who are in active therapy, we should be communicating at least once a week.”

As the EOM kicks off in July 2023, Russo noted that over time, risk-based models will become commonplace in oncology and insights gained from practices that partici- pate in models such as the OCM and EOM will give rise to whole-patient care. “The landscape continues to change in oncology from a reimbursement, patient management, and popu- lation management standpoint,” Russo said. “It’s not just about curing cancer anymore... When you start looking at social determi- nants of health and managing a population across their entire journey, it really becomes about healing lives. That’s a different mind- set shift for a lot of places. Whether we make that shift now because of EOM or make it later, because the whole market is moving in that direction, it’s important to start thinking in that mind-set.”

Reference

1. Enhancing Oncology Model. Centers for Medicare & Medicaid Services. Updated May 16, 2023. Accessed May 22, 2023. https://innovation.cms.gov/innovation-models/ enhancing-oncology-model
2. External FAQs Enhancing Oncology Model. Centers for Medicare & Medicaid Services. Accessed May 22, 2023. https://innovation. cms.gov/media/document/eom-faqs
3. Enhancing Oncology Model (EOM) implementing ePROs. Centers for Medicare & Medicaid Services. Accessed May 22, 2023. https://innovation.cms.gov/media/document/eom- epros-fs
4. Cardinal Health adds new capabilities to Navista TS in advance of the Enhancing Oncology Model. Cardinal Health. March 2023. Accessed May 22, 2023. https://www.cardinal- health.com/en/services/specialty-physician-practice/resourc- es/oncology-next/technology-innovation/navista-ts-adds-new-ca- pablilites.html