A Retrospective on the Oncology Care Model

Flatiron Health recently conducted a retrospective review of the Oncology Care Model, discovering what's working, what isn't, and what this could mean for the future.

In software development, there’s a concept of a “retro.” Short for a retrospective, the process asks the engineering and product teams to review what’s been working, what hasn’t, and what they’ll commit to improving in the future. As we near the 2-year mark of the Oncology Care Model (OCM),1 we’ve had a chance to reflect on the results and reactions from the first performance period. The completion of the first performance period cycle presents an opportune time to step back to do a retro on how the model has reshaped participating practices and influenced the technology developers supporting them.

The OCM now covers 150,000 Medicare beneficiaries in 200,000 to 300,000 episodes per year,2 and we believe the sustained success and potential expansion of the model can best be driven by incremental iteration that reflects ongoing stakeholder feedback. The OCM, or any future model, will continue to have significant implications for technology developers and participating practices, as requirements become increasingly complex, practices try to mutually optimize quality and cost, and risk-sharing decisions (or obligations) loom. There’s no doubt the model has accelerated the transition from volume to value, particularly in more agile community practices, and its innovation will continue to serve as the foundation of newly proposed models. These include the Making Accountable Sustainable Oncology Networks model,3 which has been submitted to the Physician-Focused Payment Model Technical Advisory Committee.4

Although oncologists may feel they’re straddling 2 divergent payment systems, only 1 is here to stay. HHS Secretary Alex Azar recently pointed out, in remarks to US hospitals, that the value-based transformation of our healthcare system

What's Working

is one of his department’s top priorities.5 Given this, let’s look at what’s been working, what hasn’t, and what could improve in the near term for the OCM specifically and value-based care more broadly.To kick off the retro process, let’s start with reflecting on the successes of the OCM thus far, namely greater collaboration with CMS, investment in foundational infrastructure, and the implementation of care processes that can be consistently measured and adapted.

While the program continues to evolve with each consecutive reporting period, CMS has worked to cultivate collaborative partnerships, both with oncology practices and electronic health record (EHR) developers. As oncologist Barbara McAneny, MD, president-elect of the American Medical Association, said at the Association of Community Cancer Centers Annual Meeting in March,6 CMS’ constant communication and willingness to incorporate feedback from practices is fostering a more collaborative relationship with oncologists. Through their monthly EHR vendor calls or regular “Support Lunch Hours” sessions with practices, CMS has built and maintained an entire apparatus to seek program feedback and consider program adjustments. This willingness to not only seek, but also implement, feedback has materialized most recently with the adaptation of quality measure criteria (eg, the removal of the OCM-7 measure) and the adjustment of patient-level clinical data submission requirements. This partnership among practices, EHR developers, and CMS will be imperative for the development of required technology infrastructure, to make performance more predictable, and ultimately assuage practice concerns about risk sharing.

Further, although the scope seemed insurmountable when the 160-page OCM Final Rule document was released, practices have made substantial progress within all 3 areas of foundational transformation, discussed in an earlier issue in Evidence-Based OncologyTM: program administration, care process development, and performance measurement.7 In many cases, given the OCM’s rigorous quality and infrastructure requirements, practices have created new roles that focus on the program’s administration. For example, according to Toni Perry, director of quality and regulatory affairs at Tennessee Oncology, “To ensure success in accurately reporting quality metrics, Tennessee Oncology’s Quality and Regulatory Department added 4-5 full-time data abstractor positions, specifically devoted to validating, mining, and reporting quality data.”

The Monthly Enhanced Oncology Services (MEOS) payments have also enabled practices to expand both administrative and clinical resources to their broader population. “As a practice, we have increased our investment in organizational staff resources in areas such as patient advocates and navigators in order to implement the OCM principles to both OCM and non-OCM patients,” Perry said.

Within Flatiron’s network of over 50 OCM practices, we observed that many began administering their OCM program like a clinical trial: Patients must be screened for eligibility, they must be closely monitored, and detailed clinical data must be captured and reported on a regular basis. To support this imple- mentation, Flatiron leveraged its clinical trial technology to develop a screening tool that allows practices to identify, track, and manage OCM patient episodes. From this, all reporting and EHR-linked identification flows (Figure 1).

With the ability to more precisely identify patients and manage episodes, practices and EHRs have made significant investments in optimizing and codifying necessary processes for quality measurement and care planning. These workflows range from the familiar, like capturing pain, to the new and complex, like ensuring each provider referral loop is closed. EHR developers have deployed technology to make these processes as measurable, actionable, and frictionless as possible. For example, Flatiron’s OCM practices are able to quantify and track screening for pain and depression directly from clinical workflows. Similarly, practices have been able to standardize and streamline delivery of the required care plan to their patients. Flatiron’s care plan is autocompleted from existing data in the EHR, allowing physicians to focus more on patient engagement. Since the start of the program, Flatiron practices have delivered over 68,000 unique patient care plans.

What Needs Work

Redesigning products and workflows to enable practices to more effectively implement the OCM has the added benefit of encouraging data capture with maximum ease of use and reporting efficiency. This is evident in our network; for breast cancer alone, structured data completeness doubled less than 2 years from the start of the OCM program compared with a marginal increase in structured data completeness for non-OCM Flatiron practices (Figure 2).Although there’s been progress integrating foundational components of the OCM, the nuances of the model’s logic requirements and administrative burdens have tempered the pace of transformation, hindering a full pivot to cost-reduction initiatives.

To date, practices have focused on the program elements that are well defined though not easily achieved: identifying patients, billing and collecting MEOS, and reporting registry data. These requirements have created management duties that previously did not exist, and have significantly increased administrative burdens. In particular, practices struggle with the OCM patient identification and clinical data submission components; they have trouble tracking patients on oral oncolytics, a task compounded by inconsistent access to prescription fill data outside the purview of the clinic. This pain point will continue to be burdensome with the increasing adoption of oral therapies.

From July 2016 to July 2017, the percent of Flatiron network OCM episodes initiated by an oral therapy increased by about 10% to over 24,000 episodes (Figure 3). Besides patient identification, CMS has outlined significant requirements for data completeness. To avoid penalties, practices have invested significant resources in abstraction and data backfill, an operational burden that is now necessary but costly.

Future Implications

A recent JAMA study revealed that “administrative costs of care (activities relating to planning, regulating, and managing health systems and services) accounted for 8% in the [United States] versus a range of 1% to 3% in the other countries.”8 This is evident across Flatiron’s OCM network: Over 2 million data elements were reported for almost 41,000 OCM patients to the OCM registry during the last submission period. Based on an analysis of the timing of data entry, we estimate that over 36% of breast cancer staging data were entered into the EHR via data abstraction and backfill. While executing these processes is imperative to enable the focus on more esoteric objectives of the model, these administrative components have taken over a year to fully implement for sites and have distracted practices from understanding how to approach the components that more directly affect patient care. The reality is that while collecting, reporting, and measuring high-fidelity data is arduous, much like the adoption of an EHR, they are fundamental enablers of administering value-based models. With the significant administrative investments in the program focused on tackling the OCM on ramp, the vast majority of practices have not yet demonstrated savings below their target spend.The focal shift away from the administrative execution of the model toward higher-value activities, including gaining a deeper understanding of the financial model and exploring practice-specific opportunities to reduce costs, will be key to the OCM’s durability and ongoing practice success. This shift requires administrative and physician engagement to align on care coordination opportunities, enable productive conversations with the local healthcare ecosystem, and tailor care to each patient’s goals. Anne Marie Rainey, the compliance and quality control officer at Clearview Cancer Institute in Huntsville, Alabama, agrees. “Everyone at the practice—from the concierge to the physician board—needs to be aware of the program and future implications. Through the use of data-driven tools, detailed clinical reporting, and individual education we are beginning to notice shift in physician engage- ment. We’re now investing in more clinical staff and supportive care services to better address the needs of our patient population and provide higher quality, patient-centered care.”

The OCM, among other quality programs, has shown that point-of-care technology is where and how clinicians experience alternative payment models (APMs) and can be an effective means to engage physicians. They’re where policy is translated into technology that is used to manage patients in value-based models. As such, EHRs must be held accountable for doing their part to reduce the administrative and bureaucratic friction that clinicians experience so they can better spend their time optimizing patient outcomes. During the first year of the model, our clinical team received significant feedback on the cumbersome nature of diagnosing conditions and staging patients in the EHR. Shortly after, the team employed a physician-centric approach when conducting user research with doctors and redesigned the interface to ensure these workflows were intuitive, user friendly, and encouraged structured data capture for programs like the OCM. Now it takes less time to diagnose a patient’s disease, clinicians see consolidated clinical information, and content is more easily updatable for when standards or reporting requirements inevitably change.

However, developing more physician-centric products to enhance usability isn’t enough. Point-of- care data products that leverage the scale of cancer networks must be developed with the data that’s being so meticulously captured for these models. For every patient with cancer who walks into a clinic, a cohort of the most similar patients in a network, and their treatments and outcomes, could be generated. These types of predictive cohorts could enable physicians

to make more personalized treatment decisions by learning from the experience of every patient with cancer. Outcomes of similar patients with cancer could become the evidence required to make a more value-based decision or the stories physicians use to have difficult end-of-life discussions with their patients.

The onus will continue to be on EHR developers to gain an empathetic understanding of users and build intuitive products they need to succeed in this new paradigm. Meanwhile, policymakers must see healthcare technology providers not merely as secondary constituents or APM facilitators, but rather as key stakeholders who must be engaged early and often throughout model design and implementation.

In examining the OCM through the lens which engineering and product teams use to reflect on their performance, the retro, it becomes clear just how much value-based care itself is a team effort. It starts with the care team and goes far beyond, relying upon novel partnerships between payer and provider, government and health information technology, and hospital and clinic. As we enter the second, and more trying, half of the OCM, we’ll soon know if this experiment is a collective win: if patients with cancer across the country have access to higher-quality, more affordable care.

Author Information

All authors are affiliated with Flatiron Health, a healthcare technology and services company focused on accelerating cancer research and improving care based in New York, New York. Ryan Holleran is a manager of product marketing and strategy; Arif Gilani is a manager of product operations; Abigail Orlando is a data insights engineer; and Brenton Fargnoli, MD, is the medical director of value-based care and director of product marketing and strategy.References:

  1. Oncology Care Model. CMS website. innovation.cms.gov/initiatives/ Oncology-Care. Updated April 5, 2018. Accessed April 11, 2018.
  2. Kline R, Adelson K, Kirshner JJ, et al. The Oncology Care Model: perspectives from the Centers for Medicare Medicaid Services and participating oncology practices in academia and the community. Am Soc Clin Oncol Educ Book. 2017;37:460-466. doi: 10.14694/EDBK_174909.
  3. McAceny B. MASON - Making Accountable Sustainable Oncology Networks. Assistant Secretary for Planning and Evaluation/HHS website. aspe.hhs.gov/ system/files/pdf/255906/ProposalIOBS.pdf. Accessed April 13, 2018.
  4. Physician-Focused Payment Model Technical Advisory Committee (PTAC). ASPE/HHS website. aspe.hhs.gov/ptac-physician-focused-pay-ment-model-technical-advisory-committee. Updated March 10, 2017. Accessed April 13, 2018.
  5. Azar A. Remarks on value-based transformation to the Federation of American Hospitals. HHS website. hhs.gov/about/leadership/secretary/ speeches/2018-speeches/remarks-on-value-based-transformation-to-the-federation-of-american-hospitals.html. Published March 5, 2018. Accessed April 13, 2018.
  6. Dangi-Garimella S. The price of innovation when improving cancer care delivery. The American Journal of Managed Care® website. ajmc.com/ conferences/accc-2018/the-price-of-innovation-when-improving-can- cer-care-delivery?p=2. March 15, 2018. Accessed April 13, 2018.
  7. Fargnoli B, Holleran R, Kolodziej M. Why oncologists need technology to succeed in alternative payment models. Am J Manag Care. 2017;23(SP5):SP196-SP198.
  8. Papanicolas I, Woskie LR, Jha AK. Health care spending in the United States and other high-income countries. JAMA. 2018;319(10):1024-1039. doi:10.1001/jama.2018.1150.