Although the Oncology Care Model has policy goals of improving care quality and reducing costs, data operationalization has proven to be more complex and confusing than anticipated since its debut in 2015 from the Centers for Medicare & Medicaid Services. Despite these challenges, many OCM-participating practices have been successful and reap the benefit of value-based care.
Although the Oncology Care Model (OCM) has policy goals of improving care quality and reducing costs, data operationalization has proven to be more complex and confusing than anticipated since its debut in 2015 from the Centers for Medicare & Medicaid Services (CMS).
Despite these challenges, many OCM-participating practices have been successful and reap the benefit of value-based care. During anAssociation of Community Cancer Centersvirtual discussion, recapping one of its biannual OCM Collaborative Workshops, presenters spoke on behalf of the Sidney Kimmel Cancer Center (SKCC) at Thomas Jefferson University in Philadelphia, Pennsylvania.
SKCC implemented the OCM in 2016. Since then, its actuarial informatics team has developed the Cost and Utilization Splicer (CUSp) to avoid the problems that prevent OCM participants from sharing data with physicians and providers, including technological hurdles and staff reluctance.
“The way the OCM has been set up really focuses on smarter spending as the primary component in terms of what each practice needs to be successful in achieving,” said presenter Valerie P. Csik, MRP, CPPS. “This is before better health and improved care come into play.”
When Jefferson started receiving preliminary, baseline data from CMS, questions arose from many team leaders. “There were requests immediately for more recent data, since the baseline period covered a few years prior to the start of OCM,” said Csik, the project director of practice transformation at SKCC. “There were some challenges associated with getting actionable data into our team’s hands.”
CUSp uses an analytics infrastructure that continuously updates its internal and external data. It also uses EPIC, an electronic medical record system, to provide internal Medicare costs.
Conversely, the external Medicare Shared Savings Program data warehouse provides insight into beneficially level Medicare claims of Accountable Care Organizations (ACOs). ACOs are healthcare providers that voluntarily give coordinated, quality care to Medicare patients.
“In combining those 2 data sources, [we are able to identify] our OCM or potential OCM beneficiaries and use the CMS feedback report data to verify and validate,” Csik said.
CUSp then uses this data infrastructure, as well as the data visualization and analytics software Qlik Sense, to represent CMS data organized by subgroups, including service category, procedure, provider, network, and end-of-life data. These data are then compared across cancer type, attributed oncologist at the practice-level, and the OCM target.
“The information within the tool is most relevant for our practice,” said Csik and Jared Minetola, ASA, MAAA, in an interview with Targeted Therapies in Oncology. “But the framework can be adapted for other practices that have access to similar claims details.”
For each cancer type, the CUSp dashboard provides data representative of cost opportunities by comparing the average actual episode cost per provider and the target episode cost provided by CMS. For prostate cancer, for instance, the dashboard noted that the 2 largest drivers of costs at SKCC were prescription drugs (30.8%) and outpatient pharmacy (21.9%). The average episode cost related to prostate cancer was 34.5% above target.
All providers are also displayed on a scatter diagram, so they can see how they compare with their CMS target and other providers.
“Financial success in the OCM is really dependent on saving,” Csik said. Once cost opportunities are identified, the CUSp tool uses its cost splicer to identify what could be driving such costs, narrowing in on each provider.
For example, a pie chart acknowledges the distribution of cost per provider, recognizing the top 10 cost categories, including prescription drugs, chemotherapy, and medical rehabilitation. A second pie chart identifies the top 10 cost categories of all providers as a comparator.
“What we look to do next is look for areas of impact,” said Minetola, a senior actuarial analyst at SKCC.
During the presentation, he explained that the in-patient cost category provides the admission cost per episode within the last 30 days of a patient’s life. For example, if it is $19,262 per episode with 39 episodes documented, the total cost would be $751,218. By eliminating half of these admissions, the cost opportunity could be lowered by 40%.
“That is an example of what the data can do,” Minetola said. But he added that one of the greater challenges is determining how plausible it is to implement this within the practice.
After the data are conceptualized by CUSp, the critical next step is making them actionable, according to Csik and Minetola. “Providers are obviously busy and don’t necessarily have time to learn how to [navigate] through the tool.”
To overcome this challenge, SKCC uses meetings and other group sessionsalready organized by physicians and providers—to review the data with team leaders. Once the data are reviewed, follow-up CUSp reports and performance scorecards are sent to both the disease team and provider.
“Beyond that, there are smaller group sessions and scheduled one-on-ones with Jared to review the data in greater detail, because ultimately the goal is to present data into their hands to determine the goals they wanted to set as a team, what they thought was achievable,” Csik said.
During the presentation, speakers provided an update on their implementation of CUSp at SKCC. CUSp has been distributed to each of its provider groups, along with 2 performance scorecards on a monthly basis, and reviews are underway.
Challenges remain, however, including “competing priorities and initiatives,” which prevent adequate reviews of data and performance.