RASS Levels Up, Wins Innovation Award
2004 was the year that Facebook, Gmail, Firefox, and Bluetooth all came into our lives. It was also the year that CMS rolled out the Risk Adjustment Suite of Systems (RASS).
By the time 2021 rolled around, none of those technologies looked like they did at their introduction, except that RASS still ran on a Cobol-DB2 architecture.
The purpose of RASS is to receive, process, and store risk adjustment data. RASS plugs that data into CMS statistical models to calculate risk scores for Medicare beneficiaries. Today, 70 million beneficiaries receive scores which are necessary for CMS to correctly pay Medicare Advantage and Prescription Drug Plan providers.
While the number of beneficiaries has grown, so has technology. So in 2021, the Enterprise Systems Solutions Group (ESSG) undertook the task of overhauling RASS. The new product would be called rasOne, and the team that built it applied thoroughly modern IT tools and practices including DevOps, infrastructure-as-code, test-automation, and open-source Python code. All of this was enabled by a cloud environment.
“We previously wanted to do all these things,” said Nakia Heard, the product owner. “Now that we are in Amazon Web Services and we have a platform that’s in the cloud, the innovation is endless. The sky is the limit.”
That innovation was noticed recently, when RASS won a FedHealthIT Innovation Award, which was presented on June 7. Sponsored by FedHealthIT and G2Xchange, the award “recognizes and honors the Federal Health technology and consulting community by celebrating programs for driving innovation and results across the Department of Veterans Affairs, the Military Health System, the Department of Health and Human Services, and the Centers for Medicare & Medicaid Services.”
The new and improved RASS has created a number of efficiencies. It reduced 200,000 lines of code into 3,500 lines of Python and Pyspark code-base. That means that the scoring process has been reduced from 46 hours to 15. Setting up a model run now takes two to four weeks instead of three to four months, greatly eliminating labor hours.
In addition to being simpler, the new code is decoupled from business logic, modular, and can work with different types of input data.
“The original architecture was monolithic, difficult to maintain with little reusability, project manager Tom Wolfe explained in the award submission. “Business logic was embedded with thin program code causing significant effort to implement enhancements to address evolving business needs.”
One benefit of modularized code is that CMS analysts do not need any coding expertise to feed data into the scoring system. This makes rasOne much easier to operate and maintain in addition to being more efficient.
“Over the years, we’ve made improvements in process and processing time, including moving to AWS,“ says Heard. “However, with rasOne, we innovated to deliver significant value to our business owners. I’m so proud to be part of the RASS team.”