Written by Dave Nyczepir
The Centers for Medicare and Medicaid Services (CMS) want to evolve their Medicare fraud detection platform into a largely unsupervised real-time machine learning (ML) platform, according to the director of IT marketplace.
Marc Richardson said the currently supervised ML platform is able to identify suspicious behavior on the part of insurance agents and brokers, such as gambling responses on consumer apps to increase commissions or redirect the customer. consumer mail to deal directly with the government.
CMS, which is a unit of the Department of Health and Human Services, administers Medicare and also works to identify cases of fraud and abuse within the Medicare system. The department also works with state governments to administer Medicaid.
Millions of people buy health insurance on HealthCare.gov, and tens of thousands of agents and brokers are registered to help with registrations, but can be “less than ethical” in the way they present information to people. consumers and CMS, said Richardson.
“What we’ve been able to do since moving to the cloud is really leveraging [Amazon Web Services] native services and take all the transactional data that we see coming from those agent-to-broker service channels and really look for that kind of abnormal behavior, ”Richardson said at the AWS Summit in Washington, DC on Tuesday.
Before migrating to the cloud, CMS did not have the compute resources to create environments and run analysis on transactional data in real time.
Recognizing the need to be responsive, CMS wants to reduce the time it takes to detect and respond to fraud to days with an unsupervised ML platform that alerts employees when events occur. But the algorithms are not there yet.
“At some point we’re going to be limited by humans,” Richardson said. “I think we’re still not at the point where we can allow the machines to draw any conclusions. “
About four years ago, CMS started improving direct enrollment, or as Richardson calls it: “TurboTax for people signing up for coverage.”
CMS has certified third-party companies to create and manage their own front-ends, while their backends communicate with the agency’s eligibility registration system application programming interfaces to register more consumers.
By allowing private capital to be spent to capture consumers, improving direct listing “extends the reach of government,” Richardson said.
ML helps create performance engineering models to ensure third parties can log into the federal system and is also used to detect abnormal behavior of agents and brokers.